Cooperative Work and Coordinative Practices Contributions to the Conceptual Foundations of Computer-Supported Cooperative Work (CSCW) Kjeld Schmidt Part I and III Springer, London, 2011 For Irene Contents Preface ......................................................................................................................5   Part I: Progress report ..............................................................................................9   1. Cooperative work and coordinative practices ...........................................11   Part II: Surveying the connections .........................................................................39   2. Riding a tiger, or CSCW (1991) ...............................................................41   3. Taking CSCW seriously (1992) ................................................................57   4. The organization of cooperative work (1994) ..........................................87   5. Coordination mechanisms (1996) ...........................................................107   6. Of maps and scripts (1997) .....................................................................149   7. The critical role of workplace studies in CSCW (2000) .........................167   8. The problem with ‘awareness’ (2002) ....................................................175   9. Remarks on the complexity of cooperative work (2002)........................187   10. Ordering systems (2004) .......................................................................223   Part III: CSCW reconsidered ...............................................................................277   11. Formation and fragmentation................................................................279   12. Frail foundations ...................................................................................385   13. Dispelling the mythology of computational artifacts ...........................417   References ............................................................................................................441   Index ....................................................................................................................483   Preface This book is about cooperative work and the coordinative practices through which order in cooperative work is accomplished. The development of computing technologies have from the very beginning been tightly interwoven with the development of cooperative work. Indeed, in important respects the challenges facing cooperative work in different domains have at various points been decisive in motivating and shaping crucial computing technologies such as interactive computing and networking. Over the last couple of decades computing technologies are also and increasingly being developed and used for coordinative purposes, as means of regulating complex activities involving multiple professional actors, in factories and hospitals, in pharmaceutical laboratories and architectural offices, and so on. The economic importance of the applications of these coordination technologies is enormous but their design often inadequate. The problem is that our understanding of the coordinative practices, for which these coordination technologies are being developed, is quite deficient, leaving systems designers and software engineers to base their system designs on rudimentary technologies. The result is that these vitally important systems, though technically sound, typically are experienced as cumbersome, inefficient, rigid, crude. The research reflected in this book addresses these very practical problems and is concerned with trying to establish — in the intermundia between the social sciences and computer science — a conceptual foundation for the research area of Computer-Supported Cooperative Work (CSCW). What is cooperative work in the first place? Is it something of which we can talk and reason sensibly? Is it a category of practice that can be observed, described, and analyzed in anything like a rigorous manner? How do the many actors engaged in this kind of practice accomplish their tasks in an orderly fashion, without succumbing to chaos? Can we distinguish classes of practices, coordinative practices, by means of which they do so? Can they be observed, described, and analyzed? How are these coordinative practices organized? How do they evolve? How do actors manage to organize routine cooperative activities? What difficulties do they face and how do they cope with them? By means of which conventions, procedures, techniques, etc. do they regulate their joint work? How are these practices facilitated by traditional technologies, from paper and pencil, forms and binders, to time tables and archives? What are the costs and benefits of such technologies? Which issues arise when such practices are computerized, when control and execution of routines and 6 Cooperative Work and Coordinative Practices schemes are transferred to computational artifacts? Can we devise computational facilities by means of which ordinary actors themselves can develop their coordinative practices, devise methods and tools for improved coordination? Those are the kinds of questions I have been trying to answer in the course of the last twenty–five years. They are not questions for which sociology has answers, because they are not questions sociology has raised. They are questions raised by the diffusion of computing technologies in cooperative work settings. ❧ The book comprises three rather different bodies of text. The bulk of the book — Part II — consists of articles written from 1991 to 2004 in which I have addressed and explored the issues and problems of cooperative work and coordinative practices in different directions. What unites these studies is a conception of cooperative work that makes it a researchable phenomenon, amenable to a technological research program. Instead of the ideological notion of ‘cooperation’ as an ethical imperative or the sociological notion of ‘cooperative work’ as coextensive with the notion of social nature of human conduct, these studies are based on a conception of cooperative work as observable relations of interdependence that are formed in response to practical exigencies but which then in turn require the development a family of equally observable coordinative practices. The purpose of assembling these articles, of which some have reached a large audience and some not, is to place them together, back to back, and thereby highlight their connectedness. In other words, the aim is to present a set of contributions to the conceptual foundations of the field of CSCW that, although it is unfinished business as far as unified conceptual framework is concerned, is nevertheless sufficiently elaborated and tested as an ensemble to be taken on: applied, extended, amended, challenged… The articles are reprinted without substantive changes. What changes I have made are these. I have deleted ‘abstracts’ from articles that had any: they are useful in journals or conference proceedings, as announcements to the busy reader, but would here be more of a distraction. I have also removed the usual but often terse acknowledgment statements. Typographical and other minor technical faults have been corrected without notice. Similarly, incomplete or faulty citations and references have been corrected, and citations and references have been reformatted to a common standard. However, although based a common approach, these studies were not in a strong sense conducted in a planned and goal-directed manner and the resulting ensemble of articles evidently exhibits inconsistencies, false starts, in addition to the inevitable repetitions. To provide the reader with an initial overview of the meandering argument, Part I contains an introduction in the form of a ‘progress Preface 7 report’. It gives a sketch of the development of the conception of cooperative work and coordinative practices and, by making the underlying research strategy explicit, serves to show how the different contributions are somehow connected. Finally, since the research represented by these articles has had, in part at least, a distinctly programmatic character — the subtitle is intended to indicate just that — the book is also an occasion to revert to where the journey started, to the issue of what CSCW is all about. Not for the sake of whipping a dead horse, but simply because the discussion is as topical as ever. In fact, as a research area CSCW is in disarray, and it is time to reconsider CSCW’s research program. This is the aim of Part III. Copenhagen, 1 May 2010 Kjeld Schmidt Acknowledgments The history of my research, as reflected in this book, is a clear demonstration of how tricky the concept of cooperative work can be. Some articles were obviously written in close collaboration with colleagues, while others, the majority, were written by myself. But even these could not have been written had I not been collaborating, in different ways, with a large number of colleagues. In fact, irrespective of the formal authorship of the individual articles, the general framework developed in the articles collected here has evolved over many years in more or less continual debates with scholars with whom I have debated my work and who have offered opposition, often staunch but always stimulating opposition, to my notions and contentions. I should mention, at the very least, Hans Andersen, Liam Bannon, Susanne Bødker, John Bowers, Geof Bowker, Giorgio De Michelis, Peter Carstensen, Eli Gerson, Christine Halverson, Christian Heath, Erling Havn, Betty Hewitt, Thomas Hildebrandt, John Hughes, Rachel Israël, Bjarne Kaavé, Finn Kensing, Kristian Kreiner, Jacques Leplat, Paul Luff, Gloria Mark, Morten Nielsen, Irene Odgaard, Wolfgang Prinz, Dave Randall, Jens Rasmussen, Mike Robinson, Tom Rodden, Yvonne Rogers, Pascal Salembier, Dan Shapiro, Wes Sharrock, Carla Simone, Susan Leigh Star, Lucy Suchman, Carsten Sørensen, Halina Tomaszewska, and Ina Wagner. As is typical of research work these days, my work has been carried out in collaboration with countless partners and coworkers in a number of European and Danish research projects such as, to name but the most important, FAOR, TIA, CoTech, MOHAWC, COMIC, COTCOS, DMM, DIWA, FASIT, IDAK, HIT, CITH, Cosmobiz… Those who were involved too will recognize the acronyms and will know my debt. It also so happens that the research reflected in this volume has been carried out while I was working for a string of institutions: Dansk Datamatik Center, the research center of the Danish Trade Union Federation (LO), Risø National Laboratory, the Technical University of Denmark, the IT University of Copenhagen, the University of Siegen, and Copenhagen Business School. Without the support of these institutions, none of this could have been accomplished. 8 Cooperative Work and Coordinative Practices An early version of the present book, carrying the same title, was submitted to the IT University of Copenhagen for the dr.scient.soc. degree. The two official opponents, Wes Sharrock and Yrjö Engeström, were gracious and the degree was awarded me in June 2007. The present book differs from the dissertation in many respects. Most importantly, it contains new chapters in which I, at the instigation of the anonymous reviewers, undertake a critical discussion of CSCW. The new chapters are the three that make up Part III. On the other hand, in order to prevent the book from becoming excessively large, two articles have been omitted from Part II. A couple of sojourns as visiting professor with Volker Wulf’s group at the University of Siegen, Germany, in 2007 and 2008 gave me the welcome opportunity to begin drafting the new chapters. A version of the account of the formation and fragmentation of the CSCW (in Chapter 11) was used as a basis for an article published as a discussion paper (‘Divided by a common acronym’) at ECSCW 2009. A planned chapter on ‘The concept of work in CSCW’ was taken out as used as a basis for another article and submitted for COOP 2010. Readers who want to inspect those aspects of my critique of the state of CSCW are referred to these articles. I was fortunate that a number of colleagues, among them Liam Bannon, Susanne Bødker, Lars Rune Christensen, Lise Justensen, Dave Randall, Satu Reijonen, Signe Vikkelsø, and Volker Wulf, have commented on versions of Part III (or, rather, fragments thereof), directing my attention to all kinds of shortcomings, not least points where the argument was acutely in need of clarification. I thank them all, hasting to add that the responsibility for remaining shortcomings in terms of style, grammar, logic, clarity, judgment, and plain good sense is mine alone. Part I Progress report ‘So here I am, in the middle way, having had twenty years […] Trying to learn to use words, and every attempt Is a wholly new start, and a different kind of failure […] And so each venture Is a new beginning, a raid on the inarticulate With shabby equipment always deteriorating In the general mess of imprecision of feeling, Undisciplined squads of emotion. And what there is to conquer By strength and submission, has already been discovered Once or twice, or several times, by men whom one cannot hope To emulate—but there is no competition— There is only the fight to recover what has been lost And found and lost again and again: and now, under conditions That seem unpropitious.’ T. S. Eliott: Four Quartets Chapter 1 Cooperative work and coordinative practices Over the last few decades, the interests and concerns of researchers from areas or disciplines that are otherwise rather disparate have been converging on at set of issues that are closely related, in practical terms as well as conceptually, and which all somehow center upon cooperative work practices. We have by now a rather overwhelming body of literature that, in different ways, is concerned with issues of cooperative work practice, although the issues, as is always the case, are named and framed differently by different research traditions: ‘articulation work’ (Strauss, 1985; Strauss, et al., 1985; Gerson and Star, 1986), ‘situated action’ (Suchman, 1987), ‘due process’ (Gerson and Star, 1986), ‘working division of labor’ (Anderson, et al., 1987), ‘actor networks’ (Latour, 1987; Law and Hassard, 1999), ‘horizontal coordination’ (Aoki, 1988), ‘boundary objects’ (Star and Griesemer, 1989; Star, 1989), ‘distributed cognition’ (Hutchins, 1991, 1995), ‘socially shared cognition’ (Resnick, et al., 1991), ‘distributed decision-making’ (Rasmussen, et al., 1991), ‘communities of practice’ (Lave, 1991), ‘coordination theory’ (Malone and Crowston, 1990, 1992), ‘cooperative work’ (Bannon and Schmidt, 1989; Schmidt and Bannon, 1992), ‘heedful interrelating’ (Weick and Roberts, 1993), ‘contextualized integration of human activities’ (Harris, 1995, 2000), ‘team work’ (Grudin, 1999), ‘team situational awareness’ (Endsley and Garland, 2000; McNeese, et al., 2001), ‘embodied interaction’ (Dourish, 2001), etc. Whatever the name and irrespective of the frame, the various research undertakings listed above focus on problems such as: How is concerted action of multiple individuals actually accomplished? Through which practices is such action coordinated and integrated? How do actors manage to act in a sufficiently concerted way, under conditions of partial knowledge and uncertainty, and how do they routinely manage to do their joint work in an orderly fashion in spite of local troubles and the heterogeneity of interests, approaches, and perspectives? What is 12 Cooperative Work and Coordinative Practices the role of formal constructs such as checklists, plans, blueprints, standard operating procedures, classification schemes, coding schemes, notations, etc.? How are they constructed, appropriated, applied, amended? What is the role of material artifacts and practices of writing in this context? How do actors interact through inscribed artifacts and how do they coordinate and integrate their individual activities by means of such devices? How do the different material characteristics of infrastructures, settings, and artifacts impact on cooperative practices? How do transformations of these material artifacts, media, and modalities affect the practices and the organization of cooperative work? These questions have also defined the research work reflected in this book, but the research strategy that has been developed and pursued in the course of this work differs in important respects from some of the other approaches. Outlining this strategy and how it has developed is the topic of this chapter. A brief account of how it all began in my individual case is an appropriate place to start. 1. The road to CSCW The problem of cooperative work — understanding the changing forms cooperative work takes under different economical and technological conditions as well as the skills involved in cooperative work — has, it seems, been with me for ages. Since I, as a young man in the late 1960s, immersed myself in Marx’ Grundrisse (1857-58b) and Das Kapital (1867b), the phenomenon of cooperative work has played a central role in my understanding of working class organization, that is, cooperative work conceived of as the material source of working class autonomy and assertiveness and as the source of a progressive constitution of modern industrial society based on the ‘association of free producers’. In fact, I became a sociologist in an attempt to understand these issues and my very first publications addressed those very issues (e.g., 1970). Then came the so-called ‘microprocessor revolution’ of the early 1980s. With the microprocessor the computer became a commodity. It became economically feasible not only to incorporate computers in plastic boxes with keyboards and screens that could be sold to individuals as ‘personal’ appliances, but also to incorporate computers in virtually any part of the production facilities in industrial settings. The world of industrial production was in for radical transformation. In industry, apart from ‘process industries’ such as chemical industries and power plants where key processes are automatic by nature, manual control had been prevalent up to this point in time. It is true that the overall flow of materials and parts had been mechanized in mass-production industries. The assembly lines of the automobile industries were of course famous but of marginal economic importance; the typical manufacturing enterprise was more like an engineering shop than a mass-production plant. And in any case, in engineering shops and largescale manufacturing alike, production control was still ‘manual’: the control of the Progress report: Cooperative work and coordinative practices 13 production processes (cutting, grinding, welding, etc.) were ‘in the hands’ of workers. Now, with the advent of the micro-processor, this began to change on a giant scale. In rapid succession an extended family of new technologies were devised and began to be rolled out: computer-controlled machining centers, industrial robots, automated materials-handling and transportation systems, flexible manufacturing systems (FMS), computer-aided design and manufacturing (CAD/CAM), production-planning and control systems (MRP), and so on (Schmidt, et al., 1984). It was also evident that similar upheavals were underway outside of production proper, in design and engineering, in the administrative domain, etc. I was thrilled. The chance of witnessing, in one’s life time, a technological, organizational, occupational, social, and economical transformation of such magnitude and scope was not lost on me. In that mood, it did not take long for me to decide to enter the fray. My move was, more specifically, motivated by my realization of a number of deep-set methodological problems with the program I had been pursuing until then. (1) A sociological study that tries to determine the social and organizational impact of a specific technological change is faced with a methodological nightmare. The reason is that, in order to do so, one must first of all be able to characterize the technology in question with respect to actual practices. How can a sociologist, of all people, adequately characterize technologies? Worse, how does he do that with novel, perhaps not yet fully developed technologies? How does he do it without understanding the specific roles of different technologies in actual working practices? What happens is of course that sociologists stick to secondguessing, that is, to produce post-festum ‘predictions’, or that they produce forecasts on an aggregate level so elevated as to be meaningless. (2) As pointed out by none other than Marx (1877, 1881), there is no ‘superhistorical’ ‘master key’ that allows us to anticipate societal developments in any specificity. Processes of social change may exhibit striking regularities but they always play out in a particular ‘historical milieu’, as a result of which they may have widely different effects in different local settings. The industrial revolution in Britain, for example, produced an entirely different ‘historical milieu’ for preindustrial branches of production in Britain as well as for other countries. Understanding technological change and its impact is no exception. Even when a specific technology is understood and has been adequately characterized, its ‘effects’ may differ widely according to the socio-economic milieu. e.g., the national and regional rate of employment, the quality and coverage of the educational system, migration patterns, social security systems, labor protection, etc. (3) Moreover, it quickly dawned on me that central to the transformation process generated by ‘the microprocessor revolution’ were some very complex research issues such as, for example, the famous one of ‘allocation of functionality between human and machine’. Enlightened engineers were beginning to realize 14 Cooperative Work and Coordinative Practices that the default strategy of automating whatever can be automated leads to all kinds of dysfunctional socio-technical systems and, especially in the case of safety-critical systems, possibly to accidents and disasters. They discovered that the ‘allocation of functionality’ is a design problem in its own right. At first, it was hoped that cognitive psychology could help out by offering something close to a check list of ‘tasks’ that are optimally best allocated to machines and humans, respectively (Jordan, 1963). It did not work out that way, of course (Kantowitz and Sorkin, 1987). The problem is a wicked one (Rittel and Webber, 1973). The design of complex technical systems presumes certain ‘job designs’ which in turn presume technical systems of certain shapes and forms. Coping with the circularity of the problem requires an understanding of the ‘dynamics’ of technical system and job design, which in turn requires an understanding of the technical, organizational, and socio-economic environment. And since modern work is cooperative work, this means that the design of complex technical systems and the shape and form of the organization of cooperative work are inexorably intertwined. As I saw it, CAD/CAM systems, FMS systems, MRP systems, office information systems, etc. were all systems by means of which workers would cooperate and also coordinate their cooperative activities. My realizing these methodological problems prompted me to make my move. Twenty years earlier, in 1965, I had dropped out of university, where I was initially studying philosophy, to become a computer programmer. My motive was partly to make a living, of course, but also partly fascination with the new technology. In 1985, I made a parallel move. I left academia to join a private research laboratory (Dansk Datamatik Center) where I became responsible for the lab’s research in office information systems. That move took me directly to the research area of CSCW, which was then just being formed. In the few years I spent at DDC, I became increasingly involved in doing field work. After some initial ‘quick and dirty’ workplace studies in various administrative organizations (e.g., a regional planning office, a standardization organization), I did fieldwork in domains as diverse as engineering design (e.g., design of cement factories), mathematical research, and portfolio management. In the course of these studies, the problem of cooperative work became far more concrete to me than it had been before 1985. In a report on Integrated Engineering Workstations from 1987, I summarized the observations I had made in my studies in a few theses on ‘forms of cooperative work’ (1987), which were first turned into a short paper that I had the opportunity to present at the first European workshop on CSCW in Athens (1988b), and then into a rather long paper that was presented at a workshop on ‘technology and distributed decision-making’ in Germany (1988c). The strategy that I, in effect, pursued in this early work exploited the rather unusual experience I had gained from doing a series of workplace studies in very different settings in quick succession. It provided me with the opportunity of ob- Progress report: Cooperative work and coordinative practices 15 serving patterns of cooperative work in a large number of settings and hence of subjecting the work in these settings to a comparative analysis. Although I was doing field work ‘for money’ and did not have much time to do a systematic comparative analysis, my understanding of cooperative work, as it emerged from these studies, differed radically from the then prevailing notions of cooperative work. I already knew from the classic study by Heinrich Popitz and his colleagues (1957) that ‘group work’ is a rare occurrence in industrial settings, the ‘group fetishism’ of especially American sociology notwithstanding. In my own studies, I did not see much ‘team work’ either, nor did I observe actors solemnly decide to ‘collaborate’ to reach a ‘shared goal’. Instead, the cooperative work arrangements, which were easily observable, came across, vividly and massively, as an entirely practical matter, motivated by highly pragmatic (but contradictory) concerns such as external requirements, operational constraints, and limited resources. What I saw were patterns of cooperative work emerging, changing, and dissolving again in response to recurring or changing requirements, constraints, and resources. This insight — that cooperative work is a ubiquitous occurrence in industrial and similar work settings, that it is not something invented by sociologists, socialpsychologists, or management consultants but is a routine practical measure to meet practical needs — was the platform from which I engaged in CSCW. It first of all provided me with a basis for defining cooperative work, and hence the scope of CSCW, and was then, in turn, instrumental in pointing to some absolutely key issues for CSCW. Instead of giving a summary of the various arguments and positions that are generally spelled out well enough, and more than often enough, in this collection, I will here concentrate on the strategy rather than the particular campaigns. However, in order to give readers who are unfamiliar with the research reported in this book a chance for following the remainder of this introduction, let me introduce a very simple example of cooperative work in everyday life. 2. The concept of cooperative work: The mundane case of moving Consider two men moving a dining table set consisting of a table and six chairs from one end of a living room to the other. They may do this for all sorts of reasons. Perhaps they agree to the very idea of moving the table and even that it should be moved to that particular location. They may have discussed different possibilities and only then negotiated a solution. In this case they may be said to have a ‘shared goal’ in the sense that a certain future state of affairs, a new location of the table set, has been stated as desirable and explicitly agreed to. ‘OK’, one of them may have said, eventually, ‘I think you’re right, let’s move it to the window.’ ‘Yes, let’s try,’ the other one said. 16 Cooperative Work and Coordinative Practices Perhaps the two men do not agree on the desirability of moving the table set at all. Perhaps one of them is merely assisting the other person, for a fee perhaps, or to return a favor, or out of sheer generosity. Whatever the reason, he does not really care very much about the location of the table set. He does not ‘share’ the other man’s ‘goal’, he’s just helping out. ‘Now, where do you want it?’, he asks. ‘Over there, by the window.’ ‘All right.’ Now, whatever the social arrangement — who wants to move the table and who acquiesces and who merely lends a helping hand to someone else’s project — actually moving the table involves a specific category of interaction. Moving the chairs is straightforward. Each of the men just picks up one chair at the time, carries it to the other end, and puts it down, and then repeats the operation until they are finished. To do this, in fact, two men are not needed. It might even be easier for one man to do it, since being two requires certain coordinative measures: they must take care not to be in each other’s way or to bump into the other, and they have to make sure that they do not put down the chairs in a location where these will become an obstacle to the other or pose an obstruction when moving the table. Carrying the table is a different kind of activity. Let us say that the table is large and heavy. It just might be possible for one of them to move it to the new location by dragging or pushing it across the floor, but that would surely damage the beautiful hardwood floor and perhaps also put excessive stress on the joints of the table. That is, by being more than one to do the job, it is feasible for them to move the table without causing damage to the table, to other things, or to themselves. Now, to move the table, each of them grabs it, lifts it, and they then carry it to its new location. But apart from these individual-sounding actions (grabbing the table, lifting it, carrying it, and putting it down again), how do they do it as a joint effort? Which interactions occur between the two men to make it happen? First of all, and very much as in the case of the chairs, they must somehow agree about what to do with the table. They also need to make initial arrangements such as, who takes which end of the table, and perhaps also the exact destination. Having sorted that out, they need to synchronize their respective actions: when to pick it up, when to start walking, the general direction in which to walk, the pace of walking, and when to put it down and at which exact spot by the window. If they do not handle this coordination well, the trivial task of moving the table may turn out to be demanding and may cause some broken furniture and even an injured back. These coordinative actions can happen in myriad ways. But basically, by holding the table in their hands, they are both immediately ‘aware’ of the state of the table: its location in space (altitude, pitch, and roll), its velocity, its weight. As soon as one of them walks slightly more briskly or slows down just a little, tilts the table to this or that side, lowers it or raises it, the changed state of the table is instantly conveyed to the other man who then has to act accordingly, by do- Progress report: Cooperative work and coordinative practices 17 ing likewise or by counter-acting. In the act of carrying the table, the two men are causally interrelated. The two men are also able to interact in other ways. They may hold the table in such a way that they can see each other’s faces; each may then be able to gather from the other man’s expression if he is having problems or what he intends to do next and may adjust his own actions to that. Each of them can talk, groan, and nod to make the partner understand his problems or intentions, but also to acknowledge that the other man’s problems or intentions have been understood, and so on. If these coordinative actions — the nodding, grunting, talking, swearing, shouting — are not sufficiently effective, any one of the two men can deliberately change the state of the table (force it in a certain direction, stop abruptly, shake it, etc.) to make a point that did not come across too well verbally or through gestures. They are thus likely to succeed. Indeed, moving a table horizontally a few meters across the floor of a room does not pose extraordinary challenges to the two men’s coordinative competencies. But taking the table through a door opening might. The narrow space of a doorway, compared to that of a room, will typically impose strict constraints on the operation. The men cannot move the table horizontally through the door opening but will have to tilt it. Perhaps the width of the doorway is less than the height of the table and perhaps the next room is a narrow corridor. In this case they may have to carry the table vertically while simultaneously turning it around its vertical axis to get the legs through. In order to do this, then, they have to take the table through a carefully choreographed sequence of spatial positions while moving it forward, through the opening, into the corridor. The more severe constraints of this task, compared to moving a table to another position in the same room, change the nature of the cooperative task considerably. Since the degrees of freedom are much fewer, the activities of two men become more ‘tightly coupled’. There is, literally and metaphorically, less leeway for each of them. They need to coordinate their individual actions much more closely. Things would be somewhat different if they are, say, moving a rolled-up carpet. When taking the carpet through the doorway they can bend it fairly easily, which may practically neutralize the constraints otherwise posed by a doorway leading into a narrow corridor. Their individual actions are therefore less tightly coupled, and they need not strive to coordinate their individual actions as closely and carefully. On the other hand, however, due to the carpet’s floppiness, its state is not as immediately apperceptible to the two men as that of the table. It may not be immediately obvious, from the state of the carpet as experienced locally, by each of the two, what the other is doing to the carpet. That is, while moving the carpet does not pose strong demands on their coordinative skills, when close coordination for some reason is called for they may have to be more verbally explicit than when moving a table. 18 Cooperative Work and Coordinative Practices Imagine, finally, an effort of moving on an entirely different scale such as, for instance, when a family is moving to a new house or a firm is relocating to a new building. In such cases, more than two persons will be involved for the simple reason that the number of items to be moved is much larger. If only two men were to do the job, the exercise could easily last for weeks or months. In the case of the family’s moving to a new home, the effort may involve friends and family or professional movers; in the case of the relocation of the firm, the effort will undoubtedly require dozens of professional movers. In any event, we would observe exactly the same practices as the ones we have just described, the lifting and carrying of chairs, tables, boxes, etc. The important difference is the scale: these actions will be happening in parallel and will be repeated multiple times. However, since many more items and many more actors are involved, what decision analysts call the ‘space of possibilities’ is vastly larger than in the case of simply moving one table and six chairs from one end of a room to the other. There simply are so many things to move and so many places things can be moved to. And, to confound the problem, there are dozens of actors who are simultaneously picking up items and moving them to other locations. In such cases we will observe specialized professional practices. Furniture, lamps, carpets, etc. will carry labels telling where they are to be put (‘Kitchen’ or ‘Room K4.55’). Boxes will have similar labels indicating not only their destination but also what they contain (‘Porcelain’, ‘Unfinished manuscripts’). The movers may also have floor plans of the building, and doors may be similarly marked with labels with inscriptions that correspond to the inscriptions on the floor plan (‘Kitchen’ or ‘Room K4.55’) and on the items that are to be taken there. If the relocation operation is a large one, one may also observe a pre–specified workflow of sorts, for example a list indicating the sequence in which items belonging to particular building sections are to be relocated (‘Monday: Ground floor; Tuesday: First floor’, etc.), in order to avoid congestion, confusion, chaos. 3. Strategic distinctions Our mundane example has already induced us to make a series of important and interlaced distinctions. (1) The first distinction is of course that of individual work and cooperative work. Whereas each of the men could move the chairs individually, they could not do so, for whatever reason, when it came to moving the table. For this task the joint effort of two men was required; when approaching and then grabbing the table, the two men entered a cooperative work arrangement. One should notice, however, that there is an important issue of scope or granularity in making this distinction in an actual case. Instead of focusing on their moving the chairs and then moving the table as separate tasks, we could just as well look at the moving of the whole dining table set as a set. Looked at this way, Progress report: Cooperative work and coordinative practices 19 it would still be a cooperative effort, since two men would be required for at least part of the effort, but it would still be composed of a range of actions, of which some (moving the chairs) were ‘loosely coupled’ and others (moving the table) ‘tightly coupled’ cooperative activities. In fact, the example we are looking at here could itself be merely a small part of a much larger cooperative relocation effort. The level of scope or granularity at which we describe it depends on the purpose of our investigation. We also distinguish the work itself, the work of moving the table set, from the secondary interactions required to coordinate and integrate the contributions of multiple individuals, for which I have adopted the term used by Anselm Strauss and his colleagues: articulation work (Strauss, 1985; Gerson and Star, 1986). (2) When we describe the cooperative activities of moving the furniture, we are applying a distinct analytical perspective. We look at the cooperative effort and the practices involved in that effort without knowing very much about the socio-economic setting in which it takes place. In fact, we do not need to know the socio-economic roles of the two men: if either or both of them are wage earners and do this for a salary, or if they live there and do it for their own benefit, or if one of them is providing neighborly help. Nor do we need to know anything about their ‘state of mind’, that is, if they are happy or not happy about the whole ordeal. In short, we can focus on and investigate cooperative work and coordinative practices as a distinct domain of practice, while leaving the socio-economic and organizational setting in the background. To be more precise, a number of distinctions are involved here (cf. Schmidt, 1994c, 2002a). There is the unfolding pattern of cooperative interdependencies and interactions, as the two men engage in the task and perform their work: as they approach the table set, pick up the chairs and carry them, one at a time, to the end of the room, and then return to the table, pick it up, and carry that too. These shifting patterns of actually enacted relationships is what I call the cooperative work arrangements. Other categories of relationship can also be distinguished, in particular the relatively stable configuration of actors for which the term work organization is normally used. The distinction is that of ‘mobilization’ versus ‘deployment’, that is, between the contingency arrangement (e.g., the particular configuration of workers with a range of skills deemed adequate to handle the tasks expected on a particular shift) and the enacted arrangement. In our example of moving, two men are enlisted in the contingency arrangement because the work to be done includes the moving of a large and heavy table; the enacted arrangements coalesce and dissipate again, as the two men first move the chairs and then, jointly, the table. Both of these perspectives are essential when looking at cooperative work, not only the enacted arrangements but also the contingency arrangement, because the shifting cooperative work arrangements play out among the members of the work 20 Cooperative Work and Coordinative Practices organization. They combine and deploy as the situation unfolds, on the basis of what is to be done, what it requires, who is ready, etc. By contrast to these perspectives, in which cooperative work is conceived of as material relationships and which are central to CSCW, there are of course other perspectives, in which cooperative work is conceived of from the point of view of the socio-economic relationships that are also played out in cooperative activities, in and through the material relationships. There is the unit of appropriation through which resources are committed and pooled and the results of the effort are allocated to the participants — the economic unit, if you will. In our case, different units may be involved: the family whose furniture it is and who will hopefully benefit from the moving about, and possibly the neighbor who may be helping, or the professional movers who in turn may be wage earners or members of a cooperative. And there are, finally, the contractual arrangements through which members of the unit regulate their diverse, partially incongruent, sometimes conflicting interests and concerns: the pizza and beer that the neighbor is due, or the contract specifying the ‘transaction’ between the family and the movers. In making these distinctions, or rather, in recognizing the different domains of discourse in which we talk differently about practical organization of work, I was (implicitly) influenced by the Marxian distinction between ‘material’ and ‘social’ relationships of human sociality (for an excellent reconstruction, cf. G. A. Cohen, 1978). However, I was also strongly influenced by neo-classical institutional economics (Williamson, 1979, 1981). This is not as unprincipled as it may seem. Williamson makes the exactly same distinction as I do here: between the relationship of interdependence in work and its immediate organization (the cooperative work arrangement and the work organization) on one hand, and on the other the contractual governance arrangements regulating ‘transactions’, the relationships of ownership and appropriation. The small but important difference is that he focuses on the socio-economic relationships of transfer of ownership (‘transactions’) and considers the cooperative work arrangements as a singularity, whereas I have shifted ‘figure and ground’ and focus on the cooperative work arrangement, pushing the socio-economic relationships to the back. This distinction — between cooperative work and the contractual settings in which it is situated — is useful for defining the ‘boundary’ between CSCW and Information Systems research and other areas of organizational IT. The distinction is not, as it is sometimes posited (e.g., Grudin, 1994, 1999), one of size (‘small groups’ versus ‘organizations’) but one of perspective. The CSCW perspective addresses IT sub specie cooperative work practices, irrespective of the institutional economics (not to mention the size) of the arrangement, whereas IS research addresses IT sub specie the socio-economic interests and motives of the actors (business models and the concomitant performance measurement and remuneration arrangements). Progress report: Cooperative work and coordinative practices 21 The distinction between cooperative work and the institutional and contractual arrangement is fundamental to my strategy. It allows us to single out ‘cooperative work’ as a distinct category of practice that can be conceived of independently of actors’ motives and interests and thus to talk fairly unambiguously about ‘interdependence’. (3) In our scenario we begin to discriminate different ‘kinds’ of relationships of interdependence. We noticed, for example, that moving chairs and moving tables and moving carpets involve interdependencies with rather different characteristics. We also noticed that moving the table across a room involves interdependencies that are distinctly different from moving the same table from one room to another. These different characteristics can be conceived of a so many kinds of complexity of cooperative work. Being interdependent in work is categorially different from being ‘interdependent’ by virtue of sharing a scarce resource, such as the road system in the morning rush-hour, or being ‘interdependent’ by virtue of sharing a budget, as one does when employed with the same company. Different rules apply and hence different practices are involved. Without the distinction, the term ‘interdependence’ is analytically useless. Thus defined, the concept of ‘interdependence’ itself plays a strategic role. First of all, it has provided a firm ground for defining cooperative work in a way that does not subscribe to notions of occult alignment of minds such as ‘shared goal’ or ‘shared understanding’.1 Such mentalist definitions invariably end up in tautologies: cooperative work is defined by a shared goal, and the criterion for ascribing a shared goal to actors is that they — well, act in concert. If cooperative work is conceived of this way, we are not really, i.e., accountably, able to speak of a cooperative effort that is not carried out successfully. By contrast, the concept of interdependence in work enables us to conceive of cooperative work in terms of actual observable conduct. In addition it has served a heuristic or methodological function. When we conceive of cooperative work in terms of observable interdependencies, the obvious next step is to investigate the different characteristics of different relations of interdependence, such as, for instance, the ‘degree of coupling’, the direction of dependence, as well as irreversibility, uncertainty, and various temporal characteristics, etc. This has proved analytically quite productive in the context of workplace studies, by offering a useful path towards a systematic conceptual framework for workplace studies and for comparative analysis. The concept of interdependence expresses the particular material, dynamic, and environmental characteristics of a particular cooperative effort. It therefore also 1 I am not exaggerating. Endsley and Jones, for example, explicitly conceive of ‘common goal’ and ‘shared understanding’ as ‘overlapping’ ‘sets’ and even go as far as to talk about ‘shared situation awareness’ and ‘shared mental models’ (Endsley and Jones, 2001). 22 Cooperative Work and Coordinative Practices implies that different relations of interdependence may have different characteristics. The relationships of interdependence that actors enter when forming a specific cooperative work arrangement to undertake a particular task pose specific ‘complexities’ for the cooperating actors to cope with that differ from those posed by a similar task in another setting or by another task. Carrying a table across a floor poses coordinative complexities of a more manageable kind than carrying the same table through a narrow door frame, whereas the relocation of an entire firm from one address to another in turn poses quite different and far less tractable complexities. Although the concept of complexity was at first introduced rather informally in my thinking about cooperative work and was not discussed at all until my ‘Remarks on the complexity of cooperative work’ (2002a), it served well as an intuitive and cogent way of expressing what motivated the widespread use of specialized coordinative practices involving coordinative artifacts. The concept of the complexity of cooperative work was also useful by implicitly highlighting those settings and practices for which CSCW technology might be most relevant and have the highest potential. (4) We have, in our fictional case, observed a variety of coordinative practices, ranging from the trivial one of moving the chairs prior to moving the table and avoiding collisions, to carrying the table in synchrony in the right direction, to the entirely different specialized practices of using standardized inscriptions to identify or categorize items and of developing work schedules. The concept of interdependence also, and this has been its most important strategic function, offers a framework for carefully and deliberately embracing the entire spectrum of coordinative practices, from actors’ effortlessly and unnoticeably aligning their own activities with those of others (in CSCW often referred to under the label ‘mutual awareness’), to actors’ methodically regulating their interdependent activities through pre-established schemes expressed in a set of rules (conventions, operational procedures) and concomitant, appropriately formatted textual artifacts (forms, taxonomies, schedules, etc.). So, instead of creating a categorical gulf between, say, ‘awareness’ and ‘workflows’, the concept of interdependencies enables us to conceive of coordinative practices as an open-ended repertoire of practices, some inconspicuously quotidian and ubiquitous, others exceedingly specialized and sophisticated. The idea that different cooperative work arrangements have to cope with interdependencies of different complexity and that they develop different coordinative practices to accomplish exactly that, gave me a handle on a vast and heterogeneous class of coordinative practices that all rely on coordinative artifacts and, be- Progress report: Cooperative work and coordinative practices 23 hind that, practices of writing. I dubbed them ‘coordination mechanisms’.2 They are massively present in modern cooperative work settings and one would have to be blind (or ideologically blinded) not to notice them. They are ubiquitous because economically vital. 4. Coordinative practices: From ‘coordination mechanisms’ to ‘ordering systems’ The concept of coordination mechanisms was developed in opposition to the then prevailing opinion in CSCW according to which IT systems cannot or should not regulate interaction. While the observation that formal organizational constructs are widely used in cooperative work was far from controversial, apprehension had grown and become widespread in the CSCW community with respect to the idea that computer systems could be successfully designed to regulate cooperative interaction by means of computational procedures, workflows, process models, etc. These misgiving were not at all groundless. Early attempts in CSCW to build systems that somehow imposed rules on cooperative interaction such as The Coordinator (Flores, et al., 1988), DOMINO (Victor and Sommer, 1989), etc. were generally perceived as failures, sometimes even by the designers themselves (Kreifelts, et al., 1991a), and a number of critical sociological studies by Lucy Suchman and others argued that such constructs, instead of determining action in a ‘strong sense’, by specifying step by step how work is actually performed, serve as ‘maps’ that responsible and competent actors may or may not consult to accomplish their work (Suchman, 1987). From the very beginning I found these interpretations of the experiences and of the field work data problematic and found the conclusions drawn from them unduly pessimistic. Fearing that the thinking that was already rapidly becoming the CSCW canon (Agre, 1990) would condemn CSCW research to a program that, devoid of sociological realism and practical relevance, posited that computational systems should simply provide a space of sorts for unregulated interaction (‘media spaces’, ‘workspaces’, ‘collaborative virtual environments’), I suggested an alternative research program (Schmidt, 1991a).3 Using Suchman’s dictum that ‘plans are resources for situated action’ as my shield, I very cautiously sketched my alternative approach: ‘models of cooperative work in CSCW systems (whether procedures, schemes of allocation of tasks and responsibilities, or taxonomies and thesauri, etc.) should be conceived of as resources for competent and responsible workers. That is, the system should make the underlying model 2 In fact, these practices were at first called ‘mechanisms of interaction’ (1994b), but having realized that the term ‘interaction’ covers about everything in social life and that the term ‘mechanism of interaction’ thus were far too sweeping, I later adopted the more modest term ‘coordination mechanisms’. 3 My critique of Suchman’s analysis of the role of formal organizational constructs was unfolded a few years later in my article entitled ‘Of maps and scripts’ (1997). (Cf. also the discussion in Chapter 12). 24 Cooperative Work and Coordinative Practices accessible to users and, indeed, support users in interpreting the model, evaluate its rationale and implications. It should support users in applying and adapting the model to the situation at hand; i.e., it should allow users to tamper with the way it is instantiated in the current situation, execute it or circumvent it, etc. The system should even support users in modifying the underlying model and creating new models in accordance with the changing organizational realities and needs. The system should support the documentation and communication of decisions to adapt, circumvent, execute, modify etc. the underlying model. In all this, the system should support the process of negotiating the interpretation of the underlying model, annotate the model or aspects of it etc.’ (Schmidt, 1991a) 4.1. Coordination mechanisms in practice I had originally begun to concern myself with work practices that somehow depend on formal constructs when I was doing my early work on office information systems for administrative work domains in the 1980s, but from about 1990, when I joined Jens Rasmussen’s group at Risø, my colleagues and I started on a systematic investigation of the phenomenon. Bjarne Kaavé was already engaged in his fascinating study of production planning and control practices in a Danish manufacturing plant (Kaavé, 1990). Our discussions and analyses of his observations played an important role in developing my understanding of interdependence and of the role of coordination mechanisms. A little later, Peter Carstensen and Carsten Sørensen did a study of a large industrial design project and were able to observe, virtually first hand, how a group of ordinary engineers developed and adopted a set of procedures and forms (e.g., a bug report form, a binder, a spreadsheet with a project schedule) in an attempt to cope with a cooperative effort that had become chaotic (Carstensen, 1994; Carstensen, et al., 1995a; Carstensen, 1996; Carstensen and Sørensen, 1996). To us this was a demonstration that coordination mechanisms cannot be reduced (Braverman style) to mere control instruments in the service of capital. They are, in some important respects at least, indispensable practical means for maintaining order under conditions of division of labor. Hans Andersen’s study of ‘change notes’ in another design organization complemented these findings (H. H. K. Andersen, 1994b). This view of coordination mechanisms — that they are essential means that members of cooperative work arrangements devise, adopt, and adapt in order to be able to manage their complex interdependencies — was later substantiated by a series of studies of ‘self-governing production groups’ in Danish industry that were carried out from 1998 onwards. Such groups are a key element in a strategy that aims at increasing the competitive power of manufacturing in high-cost Western countries by increasing operational flexibility and product quality. However, as it had been pointed out by Irene Odgaard in a study of production groups at a large Danish manufacturing company (1994), the groups were largely unable to accomplish the coordination tasks that had been delegated to them because they did not have the requisite tools to do it properly. Progress report: Cooperative work and coordinative practices 25 Inspired by this, my colleagues and I embarked on a series of studies of shopfloor production planning and control in Danish industrial enterprises that lasted from 1998 to 2002. In the first of these studies, of shop-floor planning and control in a manufacturing enterprise that was then switching from forecast-driven to order-driven production, we were able to show that the standard MRP system was far to crude to offer the required coordination support on the shop floor. On the other hand, it was evident that the models underlying the MRP system (e.g., bill of materials, routing schemes, processing schemes) were as indispensable as in the case studied by Kaavé. Our study resulted in a demonstrator prototype of a system that would exploit the models underlying the MRP system but give operators extensive power to overrule the plans generated by the MRP system, whenever they decided that the generated plans were inadequate, and our prototype would then, again on the basis of known interdependencies, try to anticipate the effects of the new plans enforced by the operators. One could say that the kind of system we sketched was an interactive MRP system (Carstensen, et al., 1999; Odgaard, et al., 1999). After that, in a subsequent and much larger research project, we launched a series of concurrent field studies in five Danish manufacturing enterprises: a shipyard, a maritime propulsion manufacturing plant, a cable manufacturer, a manufacturer of steel cabinets, and a manufacturer of electronic instruments (Carstensen, et al., 2001; Carstensen and Schmidt, 2002). This series of studies, in which members of production groups played an active role, further substantiated the line of thinking that was orienting our research: that the kinds of construct we have called ‘coordination mechanisms’ are of critical importance to actors in complex cooperative work settings; that actors build, adopt, manipulate, adapt such schemes when it is deemed useful to do so; that their ability to do so is critical to the productivity, effectiveness, and quality of their work and essential to their collective control of their daily working life; and that there may be potentially vast benefits to be gained from developing information technologies that support ordinary workers in those practices. In short, the studies served as ‘proof of concept’ for the concept of ‘coordination mechanism’ as well as a powerful reminder of the practical importance of the problem. 4.2. Understanding computational coordination mechanisms We knew of course, from the outset, that for coordination mechanisms to be truly viable the existing technological platforms were insufficient. Therefore, at the same time as these analytical and design studies were pursued, but tightly interlaced with them, another long-term research program was launched in close collaboration with Carla Simone and her colleagues (at the Departments of Computer Science at the Universities of Milano and Torino). 26 Cooperative Work and Coordinative Practices In this work, we understood coordination mechanisms as consisting of two basic and closely related elements: (a) a coordinative protocol: a set of rules pertaining to interaction (taken-for-granted ways of proceeding, established conventions, official policies, standard operating procedures); and (b) a coordinative artifact: a stable data structure expressed in a standardized graphical format. From our field work we concluded that a computational coordination mechanism should meet a set of requirements which we expressed as follows: Computational coordination mechanisms should be ‘malleable’. This has several implications. A CSCW system of this kind should enable actors to define the protocol of a new coordination mechanism and also to later redefine it, in order to be able to meet changing conditions, by making lasting modification to it. Furthermore, actors should be able to control the execution of the protocol and make local and temporary modifications to its behavior, for example to cope with unforeseen contingencies. In order for actors to be able to define, specify, and control the execution of the mechanism, the protocol should be ‘visible’ to actors at ‘the semantic level of articulation work’, i.e., it should be expressed in terms that are meaningful to competent members of the cooperative work arrangement. Moreover, to allow for incomplete initial specification of the protocol, it should be possible for actors to specify the behavior of the computational coordination mechanism incrementally, while it is being executed. And finally, we had observed that coordination mechanisms, even though they were typically developed for handling specific coordination issues, so to speak enter relationships with other mechanisms. More precisely, a particular coordination mechanism will typically be part of a wider complex of interdependent mechanisms. A change to the state of one mechanism may thus have implications for the state of another, and the propagation of state changes from one mechanism to another will therefore have to be taken care of somehow, manually or automatically (Schmidt, et al., 1995). Consequently, a computational coordination mechanism should be constructed in such a way that it can be linked to other coordination mechanisms in the wider setting (for the consolidated formulation of this conception, cf. Schmidt and Simone, 1996). In a systematic attempt to fully understand and, ultimately, meet these requirements an experimental notation was developed, i.e., a set of categories and predicates of articulation work and the rules of their combination. Ariadne, as the notation was called, was designed to enable actors to express, construct, maintain, execute, and link computational coordination mechanisms. (The work involved was extensive and resulted in a large number of publications. For a brief summary of this work, cf. Simone and Schmidt, 1998). In the development of the Ariadne notation, a considered decision was made to postpone the implementation and concentrate on developing a formal specification of its elements and on evaluating it against the requirements and scenarios derived from our field studies. This strategy was adopted, deliberately and explicitly, in order to avoid having the notation influenced, in an implicit and uncontrollable manner, by the inevitable limi- Progress report: Cooperative work and coordinative practices 27 tations of currently available implementation platforms. The formal specification showed that it was feasible to construct malleable coordination mechanisms by means of the notation. Subsequently, a ‘concept demonstration’ of the formal specification of the notation was implemented in an environment which is particularly suitable to managing relational structures and their behavior. This partial implementation established that the internal architecture of the Ariadne notation is workable (Simone, et al., 1995b). In the context of the overall strategy I have been pursuing, the concept of ‘coordination mechanisms’ has done a good job. It has provided a workable approach for CSCW research to address the realities of complex cooperative work settings and has thus offered an alternative to programs that in my view could cut no ice and which have subsequently been abandoned as viable programs. In addition, we were able to show that the recommended approach was theoretically feasible. 4.3. Coordination mechanisms reconsidered As noted above, we defined a coordination mechanism as consisting of a coordinative protocol as well as a concomitant coordinative artifact. This duality of coordination mechanisms was important for our work, for a number of reasons. It was crucial that the concept of coordination mechanism was not, as so often happens, instantly dissolved in idle metaphorical talk. It is almost a defining characteristic of present-day intellectual life that any new and interesting concept that arrives on the scene is immediately appropriated, stretched, transformed, abused, and eventually rendered practically useless. And it was obvious that there was a significant temptation to use the term to denote any type and form of convention, from dinner party etiquette to the grammar of ordinary speech. We therefore found it of critical importance to restrict the use of the term to the historically specific class of practices that have developed as an integral aspect of complex cooperative work practices, coordinative practices that are sufficiently standardized and specialized that they are complemented by standardized and specialized coordinative artifacts. Similarly, my colleagues and I did not want the vast array of artifacts that populate cooperative work settings, bug report forms as well as screw drivers, time tables as well as machining stations, to be included under the concept of coordination mechanisms. That would instantly render the concept meaningless. Artifacts ‘as such’ have nothing other than abstract materiality in common and is thus an empty notion. We wanted to address a specific class of artifacts, namely, specialized artifacts, coordinative artifacts, that have been devised to serve in a regulatory capacity in cooperative work arrangements and that, thus, are used in accordance with specific sets of rules, namely, coordinative protocols. In view of this, and for the sake of intellectual economy, it was assumed — or rather stipulated — that a ‘coordination mechanism’ is defined by having one and only one ‘artifact’. 28 Cooperative Work and Coordinative Practices This stipulation had an additional advantage. We had observed that coordination mechanisms, although devised for handling specific coordination issues, are regularly used in conjunction with other mechanisms. The ‘one artifact, one protocol’ stipulation seemed to make it relatively straightforward to identify and delimit individual coordination mechanisms and thereby to conceive of and construct well-bounded computational (models of) coordination mechanisms that then, again in a relatively straightforward way, could be combined to form complex coordination mechanisms. Now, let us then look at the costs of that strategy. For analytical purposes the concept of coordination mechanisms has serious shortcomings. Some of the shortcomings were known and had been identified from the beginning or early in the process, namely the ways in which coordinative issues of time and space were dealt with. As far as issues of time were concerned, when the Ariadne notation was devised, we were well aware that we only had a superficial understanding of the issue of time in coordinative practices. As a result, the temporal aspects of coordination could only be expressed as pre- and post-conditions, and as points in time, of course. The problem was duly noted and left for later work. As for spatial issues in coordinative practices, when we began to investigate production control systems such as MRP systems more in depth, it became clear that the coordinative protocols incorporated in these systems could not express spatial aspects of coordination, such as, for instance, limited storage space on the shop floor or in the shipping department. Again the problem was duly noted and left for later work. More seriously, the deliberate rigor was obtained at a high price. Firstly, the concept of coordination mechanisms was developed on the paradigm of workflows and thus does not support analysts in understanding, describing, or even noticing other kinds of coordinative protocols. Secondly, the ‘one artifact, one protocol’ stipulation is unduly restrictive when used analytically and may lead analysts to engage in futile analytical exercises. I will address the problems in that order. (1) The concept of coordination mechanisms was developed on the paradigm of pre-established workflows: an MRP system, a kanban system, a bug report, a project schedule, a change note. The kind of protocol we took as the exemplar has a fundamentally temporal structure: when A has done x, the task (in state x’) is to be transferred to B who then has to do y. The fact that we granted privileged status to protocols of this procedural kind was not the result of an oversight. It was obvious already then that other coordinative techniques play a crucial role in cooperative work, most importantly classification schemes. In fact, the Ariadne notation has a slot for ‘conceptual structures’, but these were only treated in a rudimentary manner, subordinate to procedural protocols. Our problem was not one of principle but one of practical expediency: we had not as yet had the opportunity to do proper studies of practices of classifi- Progress report: Cooperative work and coordinative practices 29 cation that would have enabled us to address them thoroughly, and so we simply left the problem for later. Anyway, this makes the concept of ‘coordination mechanism’ overly exclusive when used analytically. It excludes ab initio important coordinative practices, not by definition, but due to the framework’s lack of a rich and differentiated set of distinctions. It notoriously makes analysts overlook coordinative artifacts and protocols of crucial importance, for the simple reason that they are not, or apparently not, part of a coordination mechanism. In sum, then, workflow specifications (schedules, time tables, routing schemes) should be demoted from the status of paradigm to the one of a special case of coordinative artifacts and protocols. Although it may, in some cases, be relevant and appropriate to single out coordination mechanisms from the wider cluster of coordinative practices, the analyst should not be unaware that workflows can not work if actors are not also using, for example, maps, templates, location designators, etc., as well as classification schemes, nomenclatures, ranking schemes, verification and validation procedures, coding schemes, notations, etc. (2) As noted above, the ‘one artifact, one protocol’ stipulation turns out to be unduly restrictive too. On one hand, there evidently are coordinative protocols to which no coordinative artifacts are attached, at least not directly, and protocols evidently exist at multiple levels of abstraction. That this was also realized by us when the framework was developed is manifest from the fact that it was felt necessary to include ‘policies’ (e.g., Simone, et al., 1995b). ‘Policies’ were taken to be exactly that: global protocols, with no associated artifacts, that constrain the local specification and application of protocols. On the other hand, there evidently are coordinative protocols to which multiple artifacts are associated. It would be excessively pedantic or directly meaningless to divide such protocols into a range of discrete sub-protocols to correspond to the various distinct artifacts. To take but a simple example: Organizing a meeting in an organization, a design meeting, say, requires not only that a call or an agenda is issued, perhaps with an attached set of minutes of the preceding meeting, a list of participants, etc., but also a myriad of artifacts such as clocks and calendars, codes for rooms (names or numbers) and often inscriptions of these codes on doors, floor plans, etc. Although all of these artifacts of course presume and imply historically developed skills and conventions,4 it makes little analytical sense to insist, for each artifact, that the associated protocol must be discrete, nor does it for that matter make much sense to insist, in each and every case, on conceiving of the associated skills and conventions as coordinative protocols. The problem here is, on one hand, that artifacts such as clocks and calendars, besides their use for 4 For excellent accounts of the development of these ‘generic’ script-based coordinative practices, cf. the classic work by Jack Goody (1977, 1987), David Olson (1994), and Alfred Crosby (1997). 30 Cooperative Work and Coordinative Practices coordinative purposes in work settings, are used generally in modern civilization, for an infinite variety of purposes, and, on the other hand, that it leaves large white spots on the map of coordinative practices if they are left out. The coordination mechanism framework may thus engender a rather doctrinarian approach to analysis, and in fact, experience has shown that scholastic debates easily erupt as to whether a particular artifact is or is not part of a particular coordination mechanism. An alarm on the bridge of a ship, for instance? City maps in fire engines? Deployment plans? Access instructions? This conclusion is reinforced when we take into account not just clocks and calendars and maps and floor plans. Let us therefore, in accordance with the strategic aim of not granting privileged status to particular types of coordinative practice, try a fresh look at what is actually there. (3) What an analyst observes when entering a modern workplace is a plethora of coordinative artifacts. He or she will see bulletin boards, shift staffing plans, vacation plans, phone lists, shelves with dozens and dozens binders, often subdivided by markers, stacks of files on desks and shelves, in- and outboxes, archive boxes, production plans at work stations, production orders, part drawings, product specifications, etc. The analyst should not dogmatically exclude any of these from consideration. We have to understand how they are used, as a heterogeneous totality, in coordinative practices. For me the occasion for unpacking the concept of coordination mechanisms and resume the original program of embracing coordinative practices in their endlessly rich multiplicity arose when I seriously began to address the issues of classification schemes. Classification schemes, I knew from the very beginning, is a vitally important phenomenon, but also one for which I did not have anything like proper empirical foundation. I had observed its importance in my various studies, for instance, of the distribution of research papers within the mathematical community, of the handling of labor protection and tariff contract cases in trade unions, etc. In fact, it was to a large degree the realization that classification practices are hugely important in cooperative work, which motivated Liam Bannon and myself to highlight ‘common information spaces’ as a central problem for CSCW (Bannon and Schmidt, 1989; Schmidt and Bannon, 1992). A serious attempt to get a handle on the issue of classification was undertaken in the middle of the 1990s, in collaboration with Hans Andersen. The study of the design organization in a large Danish manufacturer of water pumps identified some, to us, very interesting coding practices, in particular a ‘product key’, a coding scheme that, by generating a unique, predictable, and reproducible designation for each of the about 25,000 product variants produced by the company, also and at the same time, generated a rigorous classification of the same items and of the hundreds of thousands of associated documents. The scheme had the additional advantage of being open-ended: changes to the design of a particular product vari- Progress report: Cooperative work and coordinative practices 31 ant, e.g., the introduction of new materials, say Teflon, for sealing a shaft, would be reflected in the coding scheme and thereby in the name and classification of the variant (Schmidt, et al., 1995; H. H. K. Andersen, 1997). But it was not until Ina Wagner and I, from 2001 onwards, jointly undertook a systematic analysis of coordinative artifacts based on the rich set of data she had gathered in the course of her long-term ethnographic study of architectural work practices, that substantial progress was finally made. 4.4. Ordering systems Investigating the large and heterogeneous collection of coordinative artifacts Ina Wagner had gathered in her ethnography of an architectural office, we began to understand that a wide variety of ordering principles were at play in the coordinative practices of architectural work. There were, of course, the temporal ordering of workflows (sequential order, stipulated phases, deadlines, versions). But there were also, intertwined with the workflows, a divine multiplicity of ordering principles, such as practices of validating documents (creator, status), as well as practices of identifying and naming documents, practices of association, aggregation, and classification of documents, coding schemes, notations, etc. Accordingly, we embarked on a meticulous analysis of each of the different coordinative artifacts used by the members of the architectural office, in an effort to reconstruct the practical logic and principles of ordering embodied in their different graphical formats and their interrelationships. We quickly abandoned all hope of coming back with a finite set of ordering principles. The task we set out to accomplish was, rather, to understand the logic of these practices, without forsaking the observed multiplicity of coordinative artifacts and practices. The strategy we adopted can be said to involve two moves: (1) The practical logic of multiplicity. The multiplicity of coordinative artifacts and protocols we observe in modern cooperative work settings is not accidental; nor is it an artifact of superficial analysis. The multiplicity is constitutive. We have long ago learned from Gerson and Star that ‘No representation of the world is either complete or permanent (Gerson and Star, 1986, p. 257). Coordinative protocols and the principles of ordering they incorporate are constructed to handle issues that are ‘local’ in the sense that they are limited to a specific activity, process, project, etc. or to a specific coordinative issue. They are specialized constructs. Rationality is always local rationality. This is not because the world is absurd or irrational, as if such a proposition would make sense. Nor is it because we mere mortals, constrained from ‘bounded rationality’, cannot grasp the rationality of the world, as if it would make sense to conceive of an agent with infinite rationality. Rationality is always local simply because the production of any kind of insight, knowledge, conceptualization, theory, representation, formulae, model, principle, 32 Cooperative Work and Coordinative Practices protocol, procedure, scheme, and so on requires time and effort and because time and effort in practice are limited resources. Rationality is local for reasons of practical epistemology. The fragmented rationality that we seem to observe in the multiplicity of protocols and artifacts is the result of what Bourdieu has called the economy of practical logic: ‘The economy of logic […] dictates that no more logic is mobilized than is required by the needs of practice’ (Bourdieu, 1980, p. 145). Practitioners are not in a position to indulge in unnecessary ‘logic’; to get the job done and in general ‘move on’, they have to economize on logic, consistency, and coherence. (2) The principle of historical specification.5 In making sense of the wealth of coordinative practices and their tricky interdependencies and in developing the required rather delicate distinctions, it was of paramount importance to steer well clear of what Gilbert Ryle has called the ‘intellectualist legend’ (Ryle, 1949). This legend is deeply entrenched in our thinking, it is a cornerstone of our modern mythology and constitutes the key strategic asset of cognitivism. In order to account for intelligent conduct, the intellectualist imputes occult operations of inference of a specifically intellectual nature to any kind of intelligent conduct. With respect to classification in particular, the intellectualist confounds specialized practices such as classification, which have been developed historically and rely on complex literate practices, with an organism’s ability to tacitly and immediately discriminate ordinary features of the world such as edible and non–edible things or sad and happy faces. The problem we were facing was the same problem that a decade previously had made us make the ‘one artifact, one protocol’ stipulation, namely, the urge to use concepts indiscriminately, to blur or ignore distinctions, which has gained so much impetus and become so prevalent in the course of the cognitivist movement. But instead of issuing a new stipulation, we were now able to express the criteria much more succinctly. What makes a coordinative protocol what it is (in addition to its specific coordinative function, of course), is not that it has a specific bond to an artifact, but that it is a specific kind of literate practice (that serves a specific coordinative function). In other words, coordinative practices are specialized practices that in turn presume an entire range of other, equally historically specific, literate practices. In our effort to get a grip on the specifics of the practices we were trying to understand, the work of Ludwig Wittgenstein and Gilbert Ryle again and again helped us to stave off imminent confusion. Furthermore, in our attempt to disentangle the web of interlaced semiological practices as (elements of) coordinative practices and to do so meticulously and accountably, the work of the British ‘integrational linguist’ Roy Harris has proved to be immensely valuable (cf., e.g., Harris, 1986, 1995, 2000). 5 The term ‘principle of historical specification’ was introduced by Karl Korsch (1938). Progress report: Cooperative work and coordinative practices 33 As this work progressed and matured, Ina Wagner and I, in our effort to be able to embrace the multifarious nature of coordinative practices in contemporary workplaces as exemplified in the work of architects, developed an approach in which coordinative artifacts and protocols in their infinite variety are taken as the point of departure, without any presumption that they bond or have to bond in specific ways. However, in going beyond the concept of coordination mechanisms, the concepts of coordinative artifacts and protocols were not abandoned at all. They are applied, not as mere subordinate elements of coordination mechanisms, but in an open-ended way, as observable and reportable phenomena. Similarly, the concept of the interlinkage of coordinative protocols and artifacts is not abandoned either but is rather, again, applied in an open-ended way. The emphasis is on how myriads of coordinative protocols and artifacts are related and connected in different ways and in an intricately recursive manner, and how they form more or less well-defined and more or less tightly coupled clusters. We call such clusters ‘ordering systems’ (Schmidt and Wagner, 2004).6 The concept is related to the concept of interlinked coordination mechanisms but does not grant privileged status to a certain kind of coordinative protocol and artifact, nor does it stipulate a strict pairing of the elements. The purpose is to support the analyst in embracing the motley of coordinative practices required in highly complex cooperative work settings. 5. CSCW’s radical program As this progress report comes to an end, a few words on the rationale of the entire program are called for. Why would a conceptual foundation for CSCW be needed in the first place? Why the systematically conceptual approach that has been pursued so doggedly? Why not simply build computational artifacts that can be put to good use here and now? Firstly, has the strategy led me the wrong way? Does the complexity of the research program I have been developing, as evidenced by its still unfinished state, the openness of the whole thing at this stage, indicate that the strategy is somehow deeply flawed, perhaps even mistaken? Might it not, after all, be wiser, to reconsider the alternative strategy, the one of developing technologically advanced ‘spaces’ of whatever kind for unregulated interaction? Well, that would amount to not answering the question that was asked and answering another one instead. However useful such ‘spaces’ might be for many purposes, they do not offer anything like a strategy for the complex cooperative work settings, the factories and hospitals, pharmaceutical laboratories and architectural offices, for which specialized coordinative protocols and artifacts, ordering systems, and galaxies of order6 We began using the term ‘ordering systems’ at the same time as the Nestor of systematic zoology, the late Ernst Mayr, began using it in exactly the same way (E. Mayr and Bock, 2002). 34 Cooperative Work and Coordinative Practices ing systems, are vitally indispensable. Practitioners simply develop these practices in order to somehow master the complexities of their interdependent work. Furthermore, the apparent simplicity of the ‘awareness space’ strategy, if one can call it that, is deceptive. The problem with the notion of ‘awareness’, as it is used in CSCW, is not just that it is poorly understood or that it barely has been defined. The problem is, I have slowly come to realize, that it grows out of an effort to give an explanation where none is needed.7 The notion of ‘awareness’ is used as a proxy for a mental state of some kind (‘awareness information’, or what have you) that the individual produces and that then somehow prompts the individual to adjust his or her conduct accordingly. As the word ‘awareness’ has been used in CSCW, by myself and many others, it is what Gilbert Ryle calls a ‘heed concept’ (1949, pp. 135 ff.). It does not explain a performance by reference to some occult preceding state; it characterizes it. Just as the term ‘intelligence’ cannot be used to explain smart conduct, but rather is used to characterize the conduct in question as inventive, ingenious, clever, diligent, adroit, imaginative, cunning, or whatever is meant in the context, the term ‘mutual awareness’ does not explain anything, nor does it stand proxy for an explanation, but is rather used for describing that a particular cooperative activity is successfully aligned and meshed, and that this was accomplished effortlessly and inconspicuously, without conversations, queries, commands, gesturing, emails, or other kinds of interruptive interaction. However, the term that was picked for this job, the term ‘awareness’, is not a ‘heed concept’ at all. One can be aware of things that one does not take into account (or heed) in one’s actions. The fact that I did not heed some good advice, could, but does not necessarily, imply that I was not aware of it: I could have ignored the advice for all sorts of reasons. ‘Awareness’ is an ‘attention concept’. It was probably picked to do the job for which it has been used because ‘being aware’ is close to ‘realizing’, ‘conscious of’, ‘noticing’ but still is used quite differently. Alan White, one of Ryle’s colleagues, has elaborated the difference well: ‘What one is conscious of or what one realises, one must be aware of. But one can become aware of things otherwise than by realising them and one may be aware of them even when one is not conscious of them. […] Being aware of something is entirely different from noticing something. We may become, remain and cease to be aware whereas we either notice or fail to notice. We can be continuously aware but we cannot continuously notice. “Noticing”, but not “being aware”, signifies an occurrence.’ (White, 1964) This is not the place to elaborate these distinctions.8 The point I am trying to make is that ‘awareness’ in CSCW has been used as a heed concept and that it, at the 7 I have sketched a critique of the notion of ‘awareness’ in CSCW in the article ‘The trouble with “awareness” (2002b) but that critique is merely an outline of what in my view needs to be done in this matter (cf. also, Schmidt and Simone, 2000). 8 Alan White’s work is both an excellent introduction to Ryle and an essential corrective (cf., e.g., White, 1964, 1967, 1982). Progress report: Cooperative work and coordinative practices 35 very same time, has had all the usual connotations of ‘awareness’: that one can ‘be aware’ without necessarily noticing whatever it is one is or becomes aware of and without realizing it. That is, ‘awareness’ has been doing not just one but two jobs in CSCW, the job of a heed concept and the job of an attention concept. Hence, I submit, all the confusion. ‘Awareness’ has been used in two ways: officially, to describe that somebody is acting in accordance with the situation, and, unofficially, to imply that this is accomplished because of some un– or subconscious processes or mechanisms or some particular mental state. As a result, it has made us look in the wrong direction. It made us search for a mental intermediary where none normally is. Cooperating actors mutually heed what each other is doing and do so effortlessly and without interrupting ongoing work because they (normally) know the work and hence know what the others are doing, could be doing, should be doing, would not be doing, and so on. They know the drill. Heeding what goes on is part of competent conduct. Their heeding is also effortless and seamless because work (normally) takes place in a material setting that is replete with indications of states and processes and activities that, for competent actors, (normally) makes it straightforward to assess the state of affairs and know what adjustments might be called for. Now, if this argumentation holds, this means that we, instead of searching for putative intermediate mental states, should try to identify the strategies competent cooperating actors employ to heed what colleagues are doing etc. How do they discriminate significant states, possible states, problematic states, etc.? What do they monitor for in the setting? What is ignored as irrelevant, what is taken into account? And so on. As the next step, this then leads to constructive explorations in an entirely different direction than that of ‘space’ technologies (no pun intended), namely, in the direction of finding ways in which these very specific monitoring strategies can be supported (by sensor technology, for instance). The conclusion is, then, that in order to support ‘mutual awareness’ adequately, the strategy would be to develop novel kinds of protocols, based on in-depth analyses of the perceptual and similar coordinative strategies of cooperating actors in complex settings, that among the actors convey selected indications about, for example, significant patterns of states. Ironically, then, instead of abandoning our program of developing technologies to support the coordinative protocols that have been developed in complex cooperative work settings, we should rather explore ways of complementing ‘natural protocols’ with ‘artificial’ coordinative protocols. The program of CSCW, when taken seriously, is indeed an ambitious one. As argued repeatedly over the years, and perhaps most explicitly in my remarks on ‘The critical role of workplace studies in CSCW’ (2000), the problem for CSCW is a radical one. If CSCW is to deliver on its promise and develop technologies that can support cooperative work as it exists out there, in laboratories, factories, and hospitals, etc., the field must be able to offer technologies that enable ordi- 36 Cooperative Work and Coordinative Practices nary workers to do, in computational environments, what they do now: express coordinative protocols and construct coordinative artifacts. In contrast to the coordination technologies that do exist and the systems that are being rolled out, the problem for CSCW is that of developing technologies that enable ordinary actors to construct computational devices by means of which they can (a) regulate their own complex cooperative activities more efficiently, safely, dependably, etc. and (b) at the same time control the regulation of their activities. More than that. If such means are not developed, the ongoing diffusion of information technologies in working life will undoubtedly cause all sorts of impediments and disruptions, as the ability of workers to manage their everyday coordinative concerns is eroded or lost. These prospects are already becoming reality. The proliferation of network protocols (ftp, email, http, chat, instant messaging, etc.) already creates parallel and mutually exclusive communication channels, which invariably impedes mutual awareness in ways that are strikingly analogous to classical multi-user database systems that were carefully designed to shield actors from experiencing the presence of others and thus create a setting in which actors can only monitor the activities of colleagues with great difficulty and outside of the system (Rodden, et al., 1992). This problem is not alleviated but rather compounded by rigid coordinative protocols that are incorporated in groupware systems and in group calendar systems, scheduling systems, booking systems, document management systems, CAD systems, electronic patient records, etc. The various protocols are of course specialized constructs that cannot simply be unified for all purposes. The problem is that, as it is, actors are left with little or no means of practical integration of the various protocols to get the job done, locally and temporarily — other than doing it manually. Thus, without a rigorous conceptual foundation, collaboration technologies are in risk of becoming part of the problem rather than the solution. 6. For lack of a conclusion The research represented by this book never had the character of a research project. It was never that well delimited. It was rather a research program, or even more to the point: the research primarily aimed at developing a research program. This is reflected in the subtitle of the book: Contributions to the conceptual foundations of CSCW. A conclusion in the standard sense is thus not appropriate. Anyway, this much has, after all, been achieved: (1) Cooperative work has been identified as a phenomenon we can study systematically, as a category of work practice, distinct from its organizational and socio-economic form, and irrespective of what mutual feelings of companionship actors may or may not have. That is, cooperative work practices have been made a researchable phenomenon. Progress report: Cooperative work and coordinative practices 37 (2) This in turn has cleared a path for making coordinative practices, their methods and techniques, a researchable phenomenon as well. (3) The research is grounded in investigations of specialized coordinative practices in different settings and has identified key features of these practices: the central role of coordinative artifacts and their associated protocols. It has shown that coordinative practices often – if not generally – involve entire clusters of such coordinative artifacts, ‘ordering systems’, and that there is what could be called a higher-order logic to this clustering, namely, that the same general schemes and notations are reused and recombined endlessly. (4) These investigations may also serve, in a loose way, as examples of how investigations of coordinative practices might be performed. Perhaps the main contribution of the research lies here, in offering, certainly not a paradigm, but some examples that other researchers may want to emulate, extend, develop. (5) It has finally been demonstrated that it is – in principle, at least – technically feasible to create computational environments by means of which ordinary workers, not programmers, can define and execute ‘coordination mechanisms’ in a fully distributed and flexible manner. For this purpose a notation for defining ‘computational coordination mechanisms’ was specified. However, the technical aspects of this research of course lies outside of the scope of the present book. The research program represented by the articles collected in Part II of this book has made substantial progress but is far from finished. It has only just begun. Many issues are still open, even wide open. Part II Surveying the connections ‘One difficulty with philosophy is that we lack a synoptic view. We encounter the kind of difficulty we should have with the geography of a country for which we had no map, or else a map of isolated bits. We can walk about the country quite well, but when we are forced to make a map we go wrong. A map will show different roads through the same country, any one of which we can take, though not two, just as in philosophy we must take up problems one by one though in fact each problem leads to a multitude of others. We must wait until we come round to the starting point before we can proceed to another section, that is, before we can either treat of the problem we first attacked or proceed to another. In philosophy matters are not simple enough for us to say “Let's get a rough idea”, for we do not know the country except by knowing the connections between the roads. So I suggest repetition as a means of surveying the connections.’ Wittgenstein, The Yellow Book (1934-35, p. 45) 277 Part III CSCW reconsidered ‘We shall not cease from exploration And the end of all our exploring Will be to arrive where we started And know the place for the first time.’ T. S. Eliot: Four Quartets 279 Chapter 11 Formation and fragmentation There is an old Danish maxim, befitting a nation of seafarers: ‘When there is wind, it’s time to make sail’. For CSCW, now is such a time. CSCW is unique in being not merely an interdisciplinary research field but a field of technological research that depends critically on in-depth studies of actual cooperative work practices in material settings. In accordance with this commitment CSCW has articulated and undertaken a critical examination and revision of fundamental assumptions and tenets in computing concerning socially distributed control of computationally distributed systems. At the same time, and by virtue of this commitment to development of technology, CSCW has been a major force in developing an understanding of work practices that has upset and overthrown the intellectualist and mechanistic (or ‘cognitivist’) notions and theories of orderly activities that only one or two decades ago seemed unassailable and unquestionable. In the process, conceptual frameworks and investigative strategies and techniques have been developed that help us to hone in on the ways in which mundane artifacts and clusters of artifacts are deployed and developed by practitioners. This has set a new and very high standard for rigorous analysis of actual work practices. On the other hand, however, many signs indicate general perplexity and a sentiment that the research program that has brought the field this far no longer offers reliable directions. New computing technologies are redefining the general technological matrix in which CSCW was originally formed and have spawned new areas of technological research. For example, technologies such as wireless networks (GSM, WiFi, Bluetooth), positioning technologies (GPS, RFID, etc.), sensor and actuator technologies, handheld and wearable devices, etc., have given rise to research areas next to CSCW such as ‘ubiquitous’ or ‘pervasive’ computing, while the very same technologies obviously offer great potential for one of the central issues in CSCW, namely the support of ‘mutual awareness’. Is that an indication that CSCW is or is on the way to become a thing of the past (cf., e.g. Crabtree, et al., 2005)? Similar concerns are engendered by the development of technologies such as high-level computational notations for ‘business process modelling’ (e.g., BPEL), ‘peer-to-peer’ protocols, ‘service-oriented architectures’ 280 Cooperative Work and Coordinative Practices (SOA), and so on (cf. van der Aalst, 2007).73 In the course of two decades, the technological matrix of CSCW has become extremely heterogeneous. Moreover, new application areas for ‘collaborative computing’ (broadly defined) have developed, such as, for instance, large-scale cooperative efforts in scientific communities (‘e-science’, ‘cyberinfrastructures’).74 In return, these application areas are bound to have significant influence on ‘collaborative computing’ (e.g., computational ‘ontologies’), but will undoubtedly also add to the heterogeneity of the matrix.75 These centrifugal forces are strengthened by a surge of studies of the various ways in which well-known ‘collaboration technologies’ are being used: the new patterns and styles of social interaction that reportedly can be observed among people communicating ‘on-line’ by means of what is often — awkwardly — called ‘social software’ or ‘social media’, a motley of protocols, services, and facilities ranging from simple protocols such as instant messaging and chat to blogs to ‘social networking’ services and facilities such as Facebook, Twitter, Second Life, Flickr, YouTube, MySpace, etc. These new ‘collaboration’ facilities, currently appearing in rapid succession, have enchanted sectors of the general public. It is hardly surprising, therefore, that this surge of public interest would captivate CSCW researchers as well; for it is of course quite interesting to sociologists and psychologists alike when a mass audience discovers and appropriates a technology, especially one that enables people to experience new forms of intercourse and interaction. However, the focus of this line of research is on the new forms of social life and human interaction these resources are reported to afford and engender. For CSCW, this represents a major problem shift in as much as the commitment to development of technology is receding. By focusing on the presumptive ‘effects’ of the various ‘social’ software and media designs, or on the new forms of interaction they afford, this line of research has, for all practical purposes, abandoned the ambition of contributing constructively, systematically, or simply accountably to the development of new technology. In short, this research is reactive with respect to technology and design (cf. Schmidt, 2009). In a related development, to some extent overlapping, CSCW’s focus on ‘work’ is increasingly considered a historical relic of no great importance (Crabtree, et al., 2005). On this view, it is considered of little or no import whether the envisioned ‘collaboration technologies’ are used for ordinary work in hospitals and factories or for games and gossip (for a critique, cf. Schmidt, 2010). 73 For an impression of the rich variety of collaborative computing technologies that are currently being investigated, cf. the recent conferences on Computer Supported Cooperative Work in Design (e.g., Shen, et al., 2008). (Cf. also: http://www.cscwd.org/). 74 These research programs have been outlined by the Atkins committee under the US National Science Foundation (Atkins, et al., 2003) and by the e-Infrastructure Working Group under the Office for Science and Innovation in the UK (2004). These new application areas are already the object of intense interest in CSCW (cf., e.g., Jirotka, et al., 2006; Lee, et al., 2010). 75 There is a considerable literature on ‘ontologies’. For an overview, cf. Gruber’s short article (2009). For initial studies of the actual construction of computational ‘ontologies’ by members of scientific communities cf. the studies by Dave Randall, Wes Sharrock and their colleagues (Randall, et al., 2007b). Chapter 11: Formation and fragmentation 281 The field is in flux. There is confusion as to the direction of the field. There is even confusion as to what constitutes an interesting and relevant problem or a valid contribution. The field is becoming fragmented. It is time to reappraise the field. The objective cannot be to restate or rephrase the traditional definitions of CSCW, for the fragmentation is a clear indication that the received understandings of the field’s program and its conceptual foundation are now deficient. We need to reconsider CSCW’s research program. It is customary to say that CSCW emerged in the late 1980s. And it is of course a fact that CSCW as an institutionalized research field was formed at this time. And indeed, this was no accident. CSCW formed in response to the possibilities and challenges posed by the then novel network technologies as represented by the Internet. All this is true and important. However, the problem with this way of framing CSCW historically is that it elides the central role that the challenges facing cooperative work practices have played in the development of computing technologies in general. It tempts us to think of CSCW as a field concerned with the application of an already existing technology and with the effects of that. The coupling of computing technologies with cooperative work is not a historical accident, nor is it, for that matter, a recent event. The challenges of enabling the actual formation of cooperative work relations in the first place, or of reducing the complexity of coordinating cooperative work, or even of eliminating the overhead cost of coordination by making cooperative work superfluous by automating the work process, — have been significant motivating forces in the development of machine technologies in general and of computing technologies in particular. It is therefore quite appropriate to start there. ‘Collaborative computing’, understood broadly and informally as computing technologies that facilitate, mediate, or support collaboration in general and cooperative work in particular, is not a yesterday’s child. Understood this broadly, ‘collaborative computing’ is about as old as travelling by jet airplane. However, continuities notwithstanding, ‘collaborative computing’ has undergone major shifts of technological paradigm. And for the purpose of reconsidering CSCW’s research program, identifying and articulating these paradigm shifts are a crucial first step. The aim of this chapter is to do just that. In fact, the bulk of the chapter is devoted to an outline of the major shifts in the prehistory and emergence of computing technologies, from the development of the concept and techniques of systematic division of labor in the 18th century to the development of the technologies of mechanical control in the Industrial Revolution, to the development of the ‘universal’ stored-program computer in the middle of the 20th century as a generic control mechanism, to the development of real-time computing during the Cold War as a means of facilitating large-scale cooperative work. This rather extensive discussion of technological developments and shifts that predate CSCW by decades and even centuries will show that CSCW, with benefit, can be seen an endeavor in continuation of the long tradition of development of 282 Cooperative Work and Coordinative Practices work organization and technology. More than that, the outline of technological development will sketch the background against which the practice-oriented research program of CSCW is to be understood: computing technologies offer unparalleled means of constructing machine systems that facilitate, mediate, and regulate cooperative work — unparalleled, not just in terms of speed of operation but also, and most importantly, in terms of operational flexibility and cost of construction and modification of designs. This has, in turn, made it technically and economically realistic to develop types of machine system, coordination technologies, that support cooperative work by regulating interdependent activities (workflow systems, etc.) but would have been practically impossible until about 20 years ago. Seen in this light, CSCW is not a avant-gardist fancy but, rather, a research effort that arises from practical concerns of ordinary work organizations. Having recounted the prehistory and formation of CSCW, I turn to the development of the different research programs in CSCW. I will first recount the story of the development, prior to the official inauguration of CSCW, of research programs concerned with ‘computer-mediated communication’ and ‘office automation’. By the time when CSCW formed as an institutionalized research field, these research programs had landed in a situation where they were seen as inherently flawed. Key researchers in the constituent communities realized that these research programs, in different ways, were based on assumptions about cooperative work practices that were fundamentally problematic: both conceptually and methodologically. This prompted what the philosopher of science Imre Lakatos (1970) has termed a ‘problem shift’. A research program was developed and articulated in which ethnographic and other kinds of in-depth workplace studies would play a key role in developing computing technologies that are adequate for the task of ‘supporting’ cooperative work by uncovering and conceptualizing the logics of cooperative work practices. After having outlined the main stages of this development, I will move on to discuss the accomplishments of this practice-oriented program: or rather, its patchy and tentative achievements. I will here focus on the complicated and apparently puzzling relationship between ethnography and the development of technology; the reason being that much of the confusion about CSCW’s program may be due to simplistic notions about the relationship between ethnography and development of technology in CSCW. My aim in all this is not to write a history of CSCW but to outline the space CSCW occupies in the greater scheme of things. 1. Cornerstones: The concepts of ‘practice’ and ‘technology’ Concepts are institutions. They change over time; not by fiat but as the accumulated result of their distributed use — sometimes coinciding, sometimes contradictory— in the normative activities of people. In the words of John Austin, ‘Our common stock of words embodies all the distinctions men have found worth drawing, and the connexions they have found worth marking, in the life-times of Chapter 11: Formation and fragmentation 283 many generations’ (Austin, 1961, p. 130). The concepts of ‘practice’ and ‘technology’ come with a suite of connotations and references, indeed a load of baggage, that we can only ignore at our peril, for then we do not know what we are in fact saying. That is we first of all have to briefly clarify the concepts of ‘technology’ and ‘technology development’ and their relationship to concepts such as ‘artifact’, ‘system’, and ‘design’ — and how these concepts are related to the concept of ‘practice’. In these matters, gross simplifications abound. While ‘technology’ to the practicing engineer is a complex of principles, models, concepts, and while it to the anthropologist is the material aspect of a culture, it is, to the sales manager of an electronics store, the gadgets on the shelves. This is confusing enough but can, with a little care, be handled as an instance of peaceful coexistence of diverging concepts. But with the notion of ‘technology development’ the situation is one of outright strife. For some, technology is simply the practical brother of science, with science providing the theories and models and with technology merely ‘applying’ the insights of science by ‘deriving’ technical principles etc. from scientific theories. For others, technology is not only a scientific discipline in its own right but, what’s more, the breadwinner of the family of the sciences. And for others again, technology is largely parasitic on the accumulated practical experience of ordinary workers. The first camp will refer to science-based technologies such as semiconductors and pharmaceuticals to justify their claims, while the third camp will refer to the record of several thousand years of history of technological development (agriculture and metallurgy, the wheel and the steam engine). Passions are running high, so high, indeed, that it is tempting to think of it all as simply different groups of social actors clamoring for honor, attention, funding, patent royalties, and a place in the sun. Anyway, in this ideological fog of scientism against romanticism against technocratic exploiters of both, the appreciation of the enormous complexity and variation of technological development is trampled flat. However, over the last few decades historians of technology have managed to develop a quite differentiated picture. Since CSCW has had its share of simplistic notions about technology development, an understanding of the complexities and variability of technology development will be essential for the purpose of reconsidering CSCW as a technological research area and of identifying the technical paradigms and program shifts that has characterized its course so far. With the concept of ‘technology’ we are fortunate in that the pedigree is fairly well documented. It originates as part and parcel of the ‘practice turn’ initiated by the intellectual movement of the 17th and 18th centuries that is generally referred to as the Enlightenment. The concept of ‘practice’, by contrast, was developed in the course of life-times of very many generations and has grown out of a tradition that goes back to early thinking about work, the nature of work, and the role of work in human life and in the very concept of humanity, and, as such, it is a tradition that has followed a course determined by both practical and intellectual con- 284 Cooperative Work and Coordinative Practices cerns. And, indeed, the modern concept of ‘practice’ was developed in intimate relationship with the development of the concept of ‘technology’. The concept of praxis (or practice) derives from antique Greek thinking of the 4th century BCE. Both Plato and Aristotle made a distinction between theoria (contemplative activity), poesis (making, production), and praxis (mere activity). While the distinction between theoria and praxis does not seem alien to modern readers, the one between poesis and praxis certainly seems odd. The distinction reflects the extremely sharp class divisions that characterized Greek society of that time.76 In his Nicomachean Ethics Aristotle stated that the concepts of ‘making [poesis] and acting [praxis] are different’, pointing out that different kinds of reasoning are involved: ‘the reasoned state of capacity to act is different from the reasoned state of capacity to make.’ He was also careful to point out that praxis is not subsumed under poesis; they are of different kinds: ‘they are not included one in the other; for neither is acting making nor is making acting.’ On the other hand, poesis involves art (techne): ‘art is identical with a state of capacity to make, involving a true course of reasoning’ (Nic. Ethics, 1140a). Now, in contrast to Plato, Aristotle did not belittle experience. It is, Aristotle remarked, ‘through experience’ that ‘science and art come to men’. This is achieved by abstraction: ‘art arises when from many notions gained by experience one universal judgement about a class of objects is produced’ (Metaphysics, 981a). In the same vein, he recognized the importance of experience in an ordinary line of action such as treating a patient: ‘With a view to action experience seems in no respect inferior to art, and men of experience succeed even better than those who have theory without experience. (The reason is that experience is knowledge of individuals, art of universals, and actions [praxis] and productions [poesis] are all concerned with the individual […])’ (Metaphysics, 981a) In the art of medicine, that is, a man that has ‘theory without experience’ is likely ‘to fail to cure’, because he does not know the individual or particular instance included in the universal, that is, because he is unlikely to be able to deal with the contingencies of the particular case. However, torn between, on one hand, a striving for knowledge of the divine and universal ‘forms’ of which particular things are but instantiations, and on the other hand a redeeming curiosity in the rich multiplicity of the world of experience, that is, trapped in an insurmountable dichotomy of the universal and the particular, Aristotle’s thinking exhibited great strain. So, having acknowledged the importance of experience for useful mundane purposes, he went on to apply the same dichotomy of the universal and the particular to his analysis of the role of experience: 76 The authority on the question of slavery in ancient Greek society is Finley (Finley, 1959, 1973). For an updated synthesis, cf. Garlan’s study (1988). For Aristotle’s view on slavery more specifically, cf. the penetrating study by Garnsey (1996). Wiedemann (1981) has produced a comprehensive collection of antique texts on slavery. — My understanding of the relationship between theoria and praxis in Aristotle in indebted to Farrington (1944/49), Redlow (1966), and Bien (1968-69, 1989). Chapter 11: Formation and fragmentation 285 ‘But yet we think that knowledge and understanding belong to art rather than to experience, and we suppose artists to be wiser than men of experience […]; and this because the former know the cause, but the latter do not. For men of experience know that the thing is so, but do not know why, while the others know the ‘why’ and the cause. Hence we think also that the master–workers in each craft are more honourable and know in a truer sense and are wiser than the manual workers, because they know the causes of the things that are done (we think the manual workers are like certain lifeless things which act indeed, but act without knowing what they do, as fire burns, – but while the lifeless things perform each of their functions by a natural tendency, the labourers perform them through habit); thus we view them as being wiser not in virtue of being able to act, but of having the theory for themselves and knowing the causes.’ (Metaphysics, 981a-b) The, for us remarkable, statement that manual workers are ‘like certain lifeless things’ that ‘act without knowing what they do, as fire burns’ is not accidental, a slip of the pen; it expresses a view with deep roots in Greek thinking, especially in the thinking of Plato and Aristotle and their schools. And it is this contempt in which they held ordinary manual work, the work of carpenters and shoemakers and ploughmen, that underlies their distinction of poesis and praxis. The slave is, Aristotle said, ‘the minister of action’ (Politics, 1254a). The underlying rationale for this categorization was first of all that the work of slaves and work that could be performed by slaves, in short, manual work, consists in bodily activities to meet bodily needs, and that whatever activity involves a significant element of bodily action, action on the condition of the material world, is slave work, praxis. His ordering can be conceived of as a scale of relations of superiority and inferiority: ‘the rule of the soul over the body, and of the mind and the rational element over the passionate, is natural and expedient’, and the same applies to men and animals, and the male and the female: ‘the one rules, and the other is ruled; this principle, of necessity, extends to all mankind’. He then went on to considering the ordering of work activities: ‘Where then there is such a difference as that between soul and body, or between men and animals (as in the case of those whose business is to use their body, and who can do nothing better), the lower sort are by nature slaves, and it is better for them as for all inferiors that they should be under the rule of a master. […] Whereas the lower animals cannot even apprehend a principle; they obey their instincts. And indeed the use made of slaves and of tame animals is not very different; for both with their bodies minister to the needs of life.’ (Politics, 1254b) Aristotle extended this to include any action that serves to satisfy a need, be it by producing necessities of life or pleasure, as opposed to knowledge that does not ‘aim at giving pleasure or the necessities of life’: ‘At first he who invented any art that went beyond the common perceptions of man was naturally admired by men, not only because there was something useful in his inventions, but because he was thought wise and superior to the rest. But as more arts were invented, and some were directed to the necessities of life, others to its recreation, the inventors of the latter were always regarded as wiser than the inventors of the former, because their braches of knowledge did not aim at utility. Hence when all such inventions were already established, the sciences which do not aim at giving pleasure or the necessities of life were discovered, and first in the places where men first began to have leisure. This is why the mathematical arts were founded in Egypt; for there the priestly caste was allowed to be at leisure.’ (Metaphysics, 981b). 286 Cooperative Work and Coordinative Practices That is, the criterion of Aristotle’s ranking of activities is the inverse of utility. Inventions aiming a giving pleasure are in turn topped by ‘the sciences which do not aim at giving pleasure or at the necessities of life’, science pursued in order to simply know. This ranking of activity and knowledge reflects two related circumstances. Aristotle took for granted that the arts that ‘aim at giving pleasure or the necessities of life’, had already completed their task. The very idea of technical development beyond what had already been accomplished, and hence the notion of building theoretical development upon practical experience, was alien to this view. This view, in turn, is intimately related to his political philosophy and his effort to perpetuate the system of slavery: ‘as the man is free […] who exists for his own sake and not for another’s’, that is, the master as opposed to the slave, ‘so we pursue this as the only free science, for it alone exists for its own sake’ (Metaphysics, 982b). Finally, as a third criterion for the ranking, Aristotle mobilized the intellectual incapacity of the slave: ‘in general it is a sign of the man who knows, that he can teach, and therefore we think art more truly knowledge than experience is; for artists can teach, and men of mere experience cannot’ (Metaphysics, 981b). Manual workers, ‘men of mere experience’, rank low, in Aristotle’s view, not because their work is deficient in some way (what they do is obviously generally successful: ‘men of experience succeed even better than those who have theory without experience’), but because they cannot explain what they are doing in terms of ‘first causes and principles’. In sum: ‘the man of experience is thought to be wiser than the possessors of any sense-perception whatever, the artist wiser than the men of experience, the master–worker than the mechanic, and the theoretical kinds of knowledge to be more of the nature of Wisdom than the productive.’ (Metaphysics, 981b-982a). And ‘the slave has no deliberative faculty at all’ (Politics, 1260a). The distinction between theoria, poesis, and praxis is an expression of this ranking scheme. Where Plato and Aristotle (and the generations of Christian scholastics and theologians who followed in their footsteps) praised bios theoretikos, the renaissance thinkers of the new and rapidly expanding world of bourgeois society had an entirely different agenda. They certainly did not subscribe to the idea that technical knowledge had achieved what could be achieved. They had things to do in this world. Frances Bacon, for example, to take perhaps the clearest voice among them, rejected the notion, received from ‘the ancients’, that anything useful at all could be accomplished when men, in ‘mad effort and useless combination of forces’ ‘endeavor by logic (which may be considered as a kind of athletic art) to strengthen the sinews of the understanding’ (Bacon, 1620, Preface). Thus, in explicit contradiction of Plato and Aristotle, Bacon argued that theory and practice are equals, so to speak, and was thereby able to even conceive of theorizing proved wrong in practice: ‘sciences fair perhaps in theory, but in practice inefficient’ (Bacon, 1620, §II:xlv): ‘Although the roads to human power and to human knowledge lie close together and are nearly the same, nevertheless, on account of the pernicious and inveterate habit of dwelling on abstractions it Chapter 11: Formation and fragmentation 287 is safer to begin and raise the sciences from those foundations which have relation to practice, and to let the active part itself be as the seal which prints and determines the contemplative counterpart.’ (Bacon, 1620, §II:iv) On this view, ordinary working practices and practical knowledge were no longer categorially separated from scientific knowledge. It was conceivable to ‘begin and raise the sciences from those foundations which have relation to practice’. However, Bacon’s ‘practice turn’ was of course largely programmatic. Production was craft-based and science immature: Galilei had just started his career when Bacon published his Novum Organon. However, a century or so later, when Denis Diderot, together with d’Alembert, edited the famous Encyclopédie, ou dictionnaire raisonné des sciences, des arts et des métiers (1751-66), the relationship between science and practice that Bacon had vaguely sensed and promulgated was becoming reality. Diderot thus wrote an article on ‘Arts’, i.e., the practical crafts, arts, techniques, and sciences, for the first volume of the Encyclopédie in which he, following Bacon, flatly observed that ‘It is man’s work [l’industrie de l’homme] applied to the products of nature’, his effort to satisfy ‘his needs’, ‘that has given birth to the sciences and the arts’ (Diderot, 1751, pp. 265 f.). He then went on to describe the relation between ‘theory’ and ‘practice’ as a reciprocal one: ‘every art has its speculation and its practice: the speculation is nothing but the idle knowledge of the rules of the art, the practical aspect is the habitual and unreflective application of the same rules. It is difficult, if not impossible, to develop the practice without speculation, and, reciprocally, to have a solid grasp of the speculation without the practice. There are in every art with respect to the material, the instruments, and the operation a multitude of circumstances which can only be learned in practice [usage]. It is for practice to present difficulties and pose phenomena, while it is for speculation to explain the phenomena and dissolve the difficulties; from which follows that hardly any but an artisan who masters reasoning that can talk well about his art.’ (Diderot, 1751, p. 266, emphases deleted). To illustrate his argument, Diderot discussed the relationship between academic geometry and the practical geometry as exercised in workshops: ‘Everyone will readily agree that there are few artists who can dispense with the elements of mathematics. However, a paradox, the truth of which is not immediately obvious, is that, in many situations, these elements would actually harm them if the precepts were not corrected in practice by knowledge of a multitude of physical circumstances: knowledge of location, position, irregular forms, materials and their properties, elasticity, rigidity, friction, texture, durability, as well as the effects of air, water, cold, heat, dryness, etc.’ (Diderot, 1751, p. 271). He went on to argue that, for instance, no levers exist ‘for which one could calculate all conditions’. Among these conditions are a large number that are very important in practice: ‘From this follows that a man who knows only intellectual [academic] geometry is usually rather incompetent and that an artist who knows only experimental geometry is very limited as a worker. But, in my opinion, experience shows us that it is easier for an artist to dispense with intellectual geometry than for any man to dispense with some experimental geometry. In spite of the calculus, the entire issue of friction has remained a matter for experimental and handicraft mathematics. […] How many awful machines are not proposed daily by men who have deluded themselves that 288 Cooperative Work and Coordinative Practices levers, wheels, pulleys, and cables perform in a machine as they do on paper and who have never taken part in manual work, and thus who never have known the difference in effect of one and the same machine in reality and as a plan?.’ (Diderot, 1751, p. 271). In other words, following Bacon, Diderot completely reversed the internal relationship of Aristotelian concept-pair ‘theoria / praxis’. When we talk of ‘practice’ we no longer conceive of it as mere regular activity devoid of ‘reasoning’ and ‘deliberation’. The notional separation of praxis and poesis has been dissolved, and both the ‘capacity of make’ and the ‘capacity to act’ have been united in the modern concept of practice — united but not conflated. The modern concept of ‘practice’ expresses and is used for emphasizing the complex dialectics of general precepts and action.77 The aim of Diderot and his fellows was not merely to celebrate of the practices of artisans and handicraft workers, but to find ways to improve received practices. The modern concept of ‘practice’ was developed as an integral intellectual component of this interventionist endeavor, as a conceptual resource for a movement devoted to understanding and transforming actual productive practices thorough investigation and rationalization. And it was in the nexus defined by the modern conception of practice in its relationship to the concepts of experience, techniques, skills, and knowledge that the concept of technology was developed. The concept of technology was developed and articulated to express and reflect the effort of developing the practices of the arts through ‘speculation’, that is, through systematic rationalization of the techniques applied in those practice. In 1675 the French minister of finance Jean-Baptiste Colbert, in his persistent effort to improve the state of the economy and promote the development of manufacturing, invited the Académie royale des sciences (founded in 1666) to produce comprehensive and detailed descriptions of the ‘arts and trades’, that is, the manifold techniques applied in the various branches of production: ‘The king wished the Academy to work unceasingly upon a treatise on mechanics, in which theory and practice should be explained in clear manner that could be grasped by everyone’ (Académie Royale des Sciences, 1729-34, p. 131). As expressed in the Academy’s own mémoire from 1699, the Academy ‘voluntarily accepted’ the assignment of ‘describing the crafts in their present condition’, knowing full well that the task was ‘dry, thorny, and not at all dazzling’. The resulting Description was to ‘penetrate to the ultimate details, although it would often prove very difficult to acquire them from artisans or to explain them’. Thereby ‘an infinity of practices, full of spirit and inventiveness, but generally unknown, will be drawn from their shadows’. In this way the crafts would be preserved for posterity, but in addition ‘able men’ who lack the leisure to visit the artisans’ workshops would be able to ‘work on the perfection’ of these practices, just as the Academy itself would not 77 Kant summarized the modern concept of ‘practice’ when he, in an essay started out by recapitulating: ‘One calls a conceptualization of rules, even of practical rules, a theory when these rules, as principles, are thought of in a certain generality and thus have been abstracted from a multitude of conditions that nonetheless necessarily influence their application. On the other hand, one does not call just any operation a praxis; rather, only such a purposive endeavor is considered a praxis that is taken to be attained by following certain generally accepted principles of procedure.’ (Kant, 1793, p. 127). Chapter 11: Formation and fragmentation 289 fail to remark if something might usefully be amended. (Académie Royale des Sciences, 1702, pp. 117 f ). This was of course not the first attempt to give in-depth accounts of specific work practices based on on-site observations.78 But with hundreds of ‘arts and trades’ to investigate and describe, the task undertaken by the Academy was of course enormous, as it involved extensive fieldwork in different lines of trade in different parts of the country. What is more, a systematic approach was required and had to be developed. It is not surprising if the research progressed only glacially. Anyway, after a couple of decades, a number of researchers (such as Billettes, Jaugeon, Carré, and others) were assigned to the task and began producing reports to the Academy, to some extent by requesting staff at the provincial prefectures to undertake the fieldwork and submit reports based on their observations. Papers were read at the regular meetings of the Academy and then filed. In 1720, the scientist René-Antoine Ferchault de Réaumur was put in charge of the cooperative effort, but by his death in 1757 only a few pieces of the accumulated analyses had been made publicly available. The reason for the lack of obvious progress — apart from the enormity of the task, of course — seems to have been dissatisfaction with the quality of the initial analytical work which was seen as not sufficiently systematic and accurate. In fact, some of the papers read at the Academy were not included in the Academy’s printed proceedings (cf. e.g., Peaucelle and Manin, 2006; Peaucelle, 2007, pp. 140 ff.). However, Duhamel du Monceau, who replaced Réaumur as project manager, succeeded in getting the publication process organized and from 1761 to 1788 altogether 81 treatises (about 100 volumes) were published under the title Descriptions des arts et métiers.79 The aim of all this, as re-stated in the Academy’s preamble to the Descriptions, was not merely to ‘examine and describe in turn all operations of the mechanical arts’ but also and equally ‘to contribute to their progress’. The Academy expected that ‘new degrees of perfection of the arts’ would be achieved when scholars undertake the effort of investigating and developing the ‘often ingenious operations performed by the artisan in his workshop; when they see by themselves the needs of the art, the boundaries at which it stops, the difficulties that prevent it from going further, the assistance that one could transfer from one art to another and which the worker is rarely expected to know.’ Subjecting work practices as they have slowly evolved from ‘obscurity’ to systematic study, rationalizing them, would show the competent worker a way to ‘overcome the obstacles that they have been unable to cross’, a way to ‘invent new tools’, etc. (Académie Royale des Sciences, 1761, pp. xvi f.). The ‘dry, thorny, and not at all dazzling’ effort of the Academy had huge impact: ‘there can be no doubt’ that contemporaneously these scientific descriptions of arts and handicrafts ‘exerted a potent influence in western Europe’ (Cole and 78 Agricola’s description of the metal trades is a case in point (1556). 79 A revised edition, collected in 20 volumes, started appearing in Neuchâtel in Switzerland shortly after (Académie Royale des Sciences, 1771-83). A German translation, also in 20 volumes, was published from 1762 to 1795 under the title Schauplatz der Künste und Handwerke (Cole and Watts, 1952, p. 18). 290 Cooperative Work and Coordinative Practices Watts, 1952, p. 1). It first of all provided a reservoir of empirical findings for much of the contents of Diderot and d’Alembert’s highly influential Encyclopédie. Although the Encyclopédie started appearing a decade ahead of the Descriptions, many of the authors who contributed to the Encyclopédie had access to the reports of the Academy researchers or were involved in both projects, so that the Encyclopédie on many topics anticipated the published contents of the Descriptions, but typically in a less detailed and accurate form.80 From the Descriptions and the Encyclopédie, the scientific description and reconstruction of the techniques of handicraft production percolated out into French and European economic life, via dictionaries and journals and other forms of popularization. Furthermore, the Descriptions provided a model for scholars that received practices were accessible to scholarly analysis and might be much improved by application of the insights, methods, etc. of the physical, chemical, mechanical, etc. sciences. Such systematic studies of work practices with a view to their rationalization were given the name ‘technology’ by the contemporary German scholar Johann Beckmann (1777).81 Referring to the Descriptions des arts et métiers and similar works as the ‘most esteemed general writings on technology’ (p. 39), that is, the model of such research, he defined ‘technology’ as follows: ‘Technology is the science of the transformation of materials or the knowledge of handicrafts. Instead of merely instructing workers to follow the master worker’s prescriptions and habits in order to fabricate a product, technology provides systematically ordered fundamental directives; how one for exactly these ends can find the means on the basis of true principles and reliable experiences, and how one can explain and exploit the phenomena occurring in the process of fabrication’ (Beckmann, 1777, p. 19) Technology, he stated, provides ‘complete, systematic, and perspicuous explanations of all works, their outcomes, and their grounds’ (p. 20). Beckmann, in his prolific research, continued the paradigm defined by the Academy in Paris. In fact, he even integrated it in his teaching, in that his students were taken to local workplaces: ‘One must have tried, without any preparation or instruction, to get acquainted with factories and manufactories to know how difficult it is’ (p. a7). One of these students, Johann Poppe, continued this work and wrote the first systematically organized history of technology.82 His brief summary of the emergence of technology as a scholarly discipline deserves quoting in this context: ‘In the eighteenth century many scholars undertook the arduous task of obtaining a precise understanding of handicraft, manufactures, and factories. Some have even made it into an specific research topic. One had become suffi80 The competition posed by the Encyclopédie may have provoked a ‘sense of urgency’ at the Academy and encouraged it to speed up it own publication effort (Cole and Watts, 1952, p. 10). 81 Beckmann (1777, p. 20) mentioned that he suggested the term with some trepidation in 1772, namely, in a review of a book by a French upholsterer on the principles of this trade: ‘Technologie oder Kenntnis der Handwerker’, or ‘technology or knowledge of the craftsman’ (1772, p. 309). (On Beckmann, cf. Exner, 1878; Beckert, 1983) 82 ‘As a history of technology of substantial scope, Poppe’s book remained almost unique for a century and a half, during which nearly everyone forgot that it existed’ (Multhauf, 1974, p. 2) — with the notable exception of Karl Marx who immersed himself in this work (Marx, 1861-63, vol. 3.1 and 3.6). Chapter 11: Formation and fragmentation 291 ciently convinced of the benefits that these efforts would have for the manual worker himself and for many managers in public office.’ (Poppe, 1807, p. 62 f.) Pointing out that the ‘stipulations and customs’ that regulate handicraft ‘were often based on deficient or even non-existent principles’ and were often influenced by ‘strong prejudices’, Poppe emphasized the role of technology: ‘Different scholars that have obtained knowledge of several handicrafts have recently, by virtue of their sciences, already cleared away many things that obstructed the greater perfection of these trades’ (ibid., p. 63), and referring explicitly to the Descriptions, he stated that ‘The meticulous description of the arts and handicrafts that the Academy brought to light belongs to the greatest scholarly works of the 18th century. It engendered several works of a similar kind, to some extent very valuable works, not only in French but also in other languages’ (ibid., pp. 91 f.). In short, the concepts of ‘technology’ and ‘practice’ were from birth joined at the hips, with technology as a systematic effort to investigate and transform the techniques applied in the practices of the useful arts. Accordingly, technology is traditionally and usefully defined as rationalized or systematic knowledge of the useful ‘arts’ or techniques (cf., Layton, 1974). Development of technology, then, is essentially a systematic conceptual endeavor that results in technical knowledge, methods, principles, etc. ‘Technology’ is an ability-word.83 The term ‘technology’ has of course been adopted for other purposes. In anthropology, for example, ‘technology’ is used as a designation for the ensemble of techniques of a particular collective of people. In short, it is used as a synonym of ‘material culture’. Now, there is certainly an important and ineluctable element of craft skill and know-how to the development and application of technology. But then there is most certainly also an important element of know-how to scientific work, as there is to all work, however ‘intellectual’ it may be. However, the problem with equating technology with technique is that this usage tends to make us blind to those practices that are specific to science and technology: systematic application of received knowledge (theoretical or empirical), rationalization of principles and methods, rigorous testing and observation, candid reporting among peers, replicability in experimental results, etc. For scientist and technologist alike, ‘similar normative imperatives remain: no engineer, anymore than a scientist, can get away with fudged data, obscure concepts, or imprecise, inadequatelydescribed measurements’ (Constant, 1984, p. 32). In short, technology, like science, is systematic knowledge. The difference lies in technology’s ‘emphasis on purpose’, as the historian of technology Donald Cardwell puts it (1994, p. 486). The technological artifact is proof that the idea actually works and works as intended. As noted by the Academy in its preamble, techniques are ‘born in obscurity’; they become technology only as a result of systematic analysis and rationaliza83 This also means that an account of technological development cannot adequately be done in the form of a list of inventions and artifacts but must be an account of development of technical knowledge, focusing on continuity and diffusion of knowledge, including conceptual fractures and problem shifts. This, of course, also applies to an account of the prehistory and formation of CSCW. 292 Cooperative Work and Coordinative Practices tion. And in fact, the great majority of inventions are and have always been what Cardwell calls ‘empirical’: inventions made ‘by arranging familiar components or materials in a novel way and without resort to abstract or scientific thinking’ (Cardwell, 1994, p. 492). Inventions that spring from craft techniques may at a later stage be subjected to reflection and undergo conceptual systematization. In fact, the development of a technology by way of systematization of empirically developed techniques have, in many cases, led to major scientific insights. Thus thermodynamics was, by an large, created by engineers and not by scientist studying heat (Cardwell, 1994). ‘Whether one takes steam power, water power, machine tools, clock making or metallurgy, the conclusion is the same. The technology developed without the assistance of scientific theory, a position summed up by the slogan “science owes more to the steam engine than the steam engine owes to science”’ (Laudan, 1984, p. 10). Now, not all technologies develop in this way: born as techniques ‘in obscurity’, which then evolve through a distributed process of incremental improvements, only to be subjected, eventually, to systematic rationalization and thus turned into a technology. Some technologies are born in the bright light of scientific insights which are then transformed into technologies: the standard example of that is of course the dramatic history of the semiconductor technology on the basis of the theories of quantum mechanics and solid state physics. Still, the transformation of articulated scientific theories and models into workable technologies is not a straightforward process of deduction; it requires a ‘separate and additional act of invention’ (Cardwell, 1994, p. 491). All kinds of practical issues have to be identified, understood, and resolved in the course of transforming scientific insight into a useful technique. As vividly illustrated by the development of semiconductor technology, theories of quantum mechanics and solid state physics only provided the essential theoretical framework of understanding the observed phenomena. To develop these insight into workable technologies, required decades of innovation work by thousands of technicians. — That is, according to Cardwell: ‘We have on the one hand crafts, technics and inventions; on the other hand technology, applied science and inventions. And common to both sides we have innovation, which we interpret as the action needed to put an invention into practice. Generally speaking, we say that inventions related to or springing from technics and crafts do not involve systematic knowledge and are, in a sense, empirical; inventions deriving from technology or applied science involve systematic or scientific knowledge.’ (Cardwell, 1994, p. 4). But, as Cardwell then points out, it is, of course, much more complicated than that: ‘there is a great deal more to effective innovation than these simple definitions suggest’ (ibid.). CSCW is often loosely described as a field ultimately devoted to the design of collaborative systems. However, this manner of speaking is misleading, for the term ‘systems design’ normally refers to the engineering practices of devising a specific configuration of existing, typically well-known, technologies (such as software architectures, protocols, modules, interfaces, etc.) so as to meet specific Chapter 11: Formation and fragmentation 293 requirements. But CSCW is not a specialized branch of practical engineering addressing the specific technical issues of designing, building, introducing, and evaluating ‘collaborative’ systems for specific settings or types of setting; it is rather a field of research devoted to the development of technologies that system designers can apply, together with other technologies, old or new, in building systems for cooperative work settings. The terms ‘technology development’ and ‘systems design’ are used for distinctly different types of socio-technical transformation. The concepts of ‘invention’ and ‘design’ are similarly categorially distinct. One could, roughly, say that ‘design’ is an intention-word, whereas ‘invention’ is an outcome-word. The concept of ‘design’ suggests premeditation, forethought in devising a plan, while ‘invention’ emphasizes the creation of something quite new. We can exhibit forethought in devising an artifact when we master the required techniques and know the odds; we can then proceed ‘by design’. When it comes to inventing something, however, we are, to some extent, in certain critical areas, fumbling, stumbling, searching for solutions. To design an artifact we rely on systematic technical knowledge; without it, we cannot exhibit forethought in devising a plan for its construction. Now, in a particular design effort there may of course be subordinate elements of invention, just as there typically are subordinate elements of design in any invention. That is, the boundary is blurred in so far as technologies are multi-level complexes. The point is that technological development provides the basis for design work, the general systematic knowledge which is then applied in the particular design solutions, while the incremental improvements represented by successive or competing design solutions in turn contributes to the further development and maturation of the technology. Technologies are, as a rule, not designed; they are developed through a series of inventions and conceptualizations. It is typically an open-ended process with all kinds of false-starts and dead-end paths, with deliberate search as well as serendipitous discoveries, with protracted periods of incremental innovation that may be interrupted by abrupt changes. The conflation of technology and systems design is rooted in the notion that technology does not fall under the broad category of knowledge but rather is a category of artifact. A candid formulation of this position can be found in George Basalla’s The Evolution of Technology (1988): ‘The artifact — not scientific knowledge, nor the technical community, nor social and economic factors — is central to technology and technological changes. Although science and technology both involve cognitive processes, their end result are not the same. The final product of innovative scientific activity is most likely a written statement, the scientific paper, announcing an experimental finding or new theoretical position. By contrast, the final product of innovative technological activities is typically an addition to the made world: a stone hammer, a clock, an electric motor. […] The artifact is a product of the human intellect and imagination and, as with any work of art, can never be adequately replaced by a verbal description.’ (Basalla, 1988, p. 30). 294 Cooperative Work and Coordinative Practices In some respects, Basalla’s argument is rather obscure, for instance when he says that ‘the artifact […] can never be adequately replaced by a verbal description’: in which sense of the words ‘adequately’ and ‘replace’ could a ‘verbal description’ possibly ‘replace’ ‘the artifact’? Is this even a meaningful proposition? What seems to be said here is anyway oddly fetishistic, in that it hinges on the implicit assumption that the product of scientific or technological work somehow speaks for itself. Basalla’s central claim here is of course that what is essential to a technology is somehow embodied in the thing itself; that being a hammer, a clock, an electric motor is somehow an intrinsic material property of the thing. But technological artifacts that are not integral to a living practice are merely a heap of junk. Or perhaps they are on exhibit in a museum as a representation of a past technology the use of which may now be unknown. In fact, it so happens modern archeology has had to develop experimental methods in order to understand the ‘real-life processes’ through which the physical remnants were produced. For example, the experimental archeologist Nicholas Toth has reconstructed the seemingly simple task of fabricating stone hammers similar to those produced by hominids living in East Africa between 2.4 and 1.5 million years ago. The research effort was extensive and exhausting, but the conclusion clear: ‘Although the products of Oldowan technology are quite simple, the processes required in the hominid mind to produce these forms show a degree of complexity and sophistication: in other words, skill’ (Schick and Toth, 1993, pp. 133 f.). In sum, ‘Technology is something much larger than the tool itself. It refers to the system of rules and procedures prescribing how tools are made and used’ (ibid., p. 49). Or in the words of the historian Rachel Laudan, the ‘mute presence of the remaining artifacts does not speak for itself’, and this is the obvious reason why ‘technological knowledge can easily be lost’ (Laudan, 1984, p. 7). I quote Basalla because his position, though deeply problematic, has been influential and is widely cited. Moreover, Basalla’s claim has been mobilized in the general area of Human-Computer Interaction (HCI) in an attempt to understand the apparently surprising, and to some unsettling, observation that the technologies of interactive computing were developed well in advance of any formulated theory of interactive computing and that the technical solutions were not ‘deduced’ from psychological theory. In 1991 John Carroll thus noted that ‘some of the most seminal and momentous user interface design work of the last 25 years’ such as Sutherland’s Sketchpad and Engelbart’s NLS in the 1960s ‘made no explicit use of psychology at all’ (Carroll, 1991, p. 1), and that ‘the original direct manipulation interfaces’ as represented by the Xerox Alto (released in 1973) and the Xerox Star (1981) ‘owed little or nothing to scientific deduction, since in fact the impact went the other way round’ (Carroll, et al., 1991, p. 79). These observations are well founded. More than that, when Carroll and his colleagues then argued that it was not embarrassing but quite legitimate for HCI to engage in post hoc ‘scientific investigations’ of the techniques of interactive computing that had developed incrementally through ‘emulation of prior art’ (ibid., p. 75), this was Chapter 11: Formation and fragmentation 295 again quite justified, for this has been and remains the normal form of technology development. However, Carroll and his colleagues then slips into conceiving of technical artifacts in the fetishistic manner propounded by Basalla: ‘it seems typical in HCI for new ideas to be first codified in exemplary artifacts and only later abstracted into discursive descriptions and principles’ (Carroll, et al., 1991, p. 79). The problem with this interpretation lies in the very notion that ‘ideas’ are somehow ‘codified’ ‘in’ ‘artifacts’. Codified? As pointed out by the molecular biologist Jacques Monod years ago, there is nothing in the form, structure, behavior of an object that makes it an artifact (Monod, 1970). What makes the thing an artifact is the practices to which it belongs, the practices for which it has been designed and built, for which it has been appropriated, and in which it is used on a regular basis, and one can only ‘decode’ the artifact if one knows and understand these practices (cf. Bannon and Bødker, 1991). After all, ‘reverse engineering’ requires the same general competencies as the engineering of a device. That is, the problem here is that, in focusing on the artifact itself, the practices as part of which the techniques developed and in which they have become integrated have been effectively expunged from consciousness. In short, technology cannot be reduced the technical artifact although the artifact, of course, plays a pivotal role in the demonstration and application of the technology. ‘Technology’ is an ability-word. To complicate matters even more, immensely more, technical knowledge has what one could call systemic character. The development of a technology will depend upon received knowledge (scientific or practical) that, although it does not provide a theory of the technology under development, is nevertheless indispensable. Thus even in the case of a technical development as conspicuously craftbased as the development of the steam engine, the mechanists who developed it would have been at a complete loss without their mastery of geometry and arithmetic and techniques of measurement. More than that, a technology typically comprises a set of what could be termed constituent technologies, that is, technologies that have been adopted, reconfigured, and put to novel uses as components of the new technology. Thus distinctly different technologies, each of which may have evolved incrementally in a distributed way, may deliberately be ‘shifted laterally’, from their context of origin, to become part of or be combined in a novel technology. A case in point is the railroad. When, in 1829, the Stephensons demonstrated their prototype steam locomotive, the Rocket, and won the competition, ‘the individual components of the railroad, considered as a system, had been assembled over the previous fifty years or so. The canal builders had mastered the techniques of drilling tunnels. building up embankments and digging out cuttings. They had established the legal precedents for compulsory purchase of land and they had learned how to muster and manage large bodies of skilled and unskilled men. The millwrights, of the mining industry had invented and developed the locomotive. The stage coach system had popularized the idea of public passenger transport at so much per mile and the discipline of the timetable. These components had been progressively refined. What was required was the ability to put them all together and weld them into the public 296 Cooperative Work and Coordinative Practices railroad so very different from the little mine trackways with their rough, primitive steam locomotives, “iron horses”. The Stephensons came to realize that the railroad could extend to working people a luxury that only the affluent had been able to afford. Their vision of a rail network to provide ordinary folk with a cheap, fast. safe and comfortable means of travelling wherever they wanted to go amounted, when combined with their practical abilities to bring it about, to the achievements of genius.’ (Cardwell, 1994, p. 234 f.) Not only were received technologies appropriated and combined in the formation of railroad technology but the accumulated experience with railroad technology prompted the deliberate development of other technologies. When railroad networks began to operate and spread, issues of coordination began to emerge and be understood. The operation of geographically dispersed but interdependent and temporally critical activities, as those of a railway service, requires communication facilities that operate at a much higher speed than the train service itself. Thus the ‘rapid spread of railroads heightened the demand for the fastest means of communication along the lines’ (Cardwell, 1994, p. 251). This practical need, to a large degree, then drove the development of the technologies of the electric telegraph. The discovery of the electromagnetic effect (by H. C. Ørsted in 1820) was by then already being applied experimentally to communication over distances of several kilometers (for research purposes) but the coordination needs of the new railroad operations in Britain soon became a major motivation for developing this technology. It was tried successfully in 1837-38, and from 1842 practical implementation got underway. By 1848, operations of about half of the British railroad lines were coordinated by telegraph (Standage, 1998, p. 61). Now, with expanding networks of rail services, coordinated by means of telegraph services, another ‘bottleneck’ emerged that required additional technological developments. The coach lines had been operating in an environment with a multitude of local times as determined by longitude, and the railroad companies initially tried to continue the practice of the coach lines. But, as the historian of technology David Landes puts it, ‘trains moved too fast for this, continually exposing passengers and crew to discrepancies and confusions’. The first step was to use the telegraph ‘to transmit almost instantaneously an exact hour and minute from the central office to every point on the line’. This established ‘a standard time for all those served by a given network’, but this of course caused coordination problems across railway networks. The next step was therefore to unify local railway practices by agreeing to adopt the local time at the Greenwich observatory as a national standard. This was done by the end of 1847. But, as Landes comments, ‘The effectiveness of the change depended, of course, on the creation of a national time service, communicating precise time signals at regular intervals to clocks and stations around the country.’ (Landes, 1983, pp. 285 f.). This was, in turn, made possible by the development of the ‘use of galvanism’, i.e., telegraphic signals, to synchronize timekeepers in different locations and, indeed, in the words of the Astronomer Royal, Airy: ‘the extensive dissemination throughout the Kingdom of accurate time-signals, moved by an original clock at the Royal Observatory’ (quoted in Howse, 1980, p. 95). Implementation of this new technique of automatic synchronization was begun on the South Eastern Railway in Sep- Chapter 11: Formation and fragmentation 297 tember 1851 and was completed by August the next year. It was in the following years extended to post and telegraph offices, town hall clocks, workshops of chronometer makers, and to factory clocks (ibid., pp. 96-105). The case of railroad technology illustrates the systemic nature of technology. A given technology presumes a — sometimes vast — network of other technologies that serve as component technologies or as part of the technical platform for the one under consideration. Sometimes the need of such auxiliary or component technologies is only understood as the new technology gets under way, problems are identified, and practitioners begin to search for solutions. And sometimes the relationship between two otherwise different technologies is such that one can conceive of their development as something akin to co-evolution: the further development of one technology depends on the progress of the other, and the advances in one technology may be held up for a long time until another technology has achieved a certain level of maturity (stability, performance/cost ratio, etc.). The systemic character of technology is a key to understanding the dramatic technological changes that sometimes occur. A new technology may be developed for certain purposes and it may then be realized by researchers or engineers that it can substitute a quite different component technology in an otherwise unrelated technological complex — and it may turn out that the substitute offers solutions to known problems or even to bottlenecks or limitations that had not yet been perceived. Sometimes the component technologies had been developed for other purposes but are then shifted laterally and combined in innovative ways to form a new technology (in the case of railroads: rails, heat engines, machine tools, time tables, etc.). The systemic character of technology has important methodological implications, in that the challenge of uncovering and grasping ‘the internal development of technology’ (Laudan, 1984) should be an overriding concern of any effort to understand the development of technologies — be it railways or CSCW. That is, it is essential to conceive of technology as systemic and hence of the research and engineering efforts of developing technical knowledge as (loosely) coherent or (tentatively) converging ‘paradigms’ (Constant, 1984; Laudan, 1984; T. P. Hughes, 1987). Before I move on, I should point out that the Kuhnian notion of ‘paradigm’ should not to be taken as equivalent to the notion of ‘theory’ but as a set of ‘universally recognized scientific achievements that for a time provide model problems and solutions to a community of practitioners’ (Kuhn, 1962, p. x), that is, a complex of notions of what constitutes a researchable, relevant, and interesting problem as well as notions of what constitutes findings and solutions; measurement standards, protocols, procedures, methods, instruments, equipment; forms of apposite argument and conceptualization as well as genres of presentation of data and findings; as well as the institutionalized forms of these notions, standards, and forms: channels, policies, and procedures of publication, review, etc.; research review boards, schemes, etc.; educational institutions, textbooks, professional organizations, and so on — all centered on and incarnated in those contributions that 298 Cooperative Work and Coordinative Practices are considered exemplary, the ‘paradigm’ (Sharrock and Read, 2002). The development of technical knowledge exhibits similar characteristics. The most important difference is, again, the ‘emphasis on purpose’ in technological research, which, in practice, is bound up with the concerns for usefulness, costs, stability, performance, maintenance, etc. There is an inherent risk in focusing on the ‘internal development of technology’, however. The concept of technology may then simply be confounded with that of engineering, and particular technologies may become categorized according to certain components technologies. This has as happened rather frequently in the historiography of computing, when ‘computing’ as a technology is dated from the Colossus (1943) and the ENIAC (1946) because they were the first electronic digital computers or the Manchester ‘Baby’ (1948) because it was the first storedprogram digital computer. They were all important steps and their design embodied technologies that have later been found important. But technology is nothing if it is not ‘a useful art’. The relationship between technology and practice is internal, that is, conceptual, like ‘figure’ and ‘ground’: you can’t have one without the other. A given technology is only a technology in relation to its application context and we can only compare and identify paradigmatic shifts in relation to that specific application context. Does this mean, then, that the ‘internal development of technology’ is a chimera? Not at all. In fact, technological development is often or, indeed, typically driven by technologists’ addressing known or anticipated internal ‘anomalies’ in the established technology, trying to overcome instabilities and performance limitations, reduce energy consumption, increase reliability, etc. The paradox dissolves as soon as we remember that technologies are systemic and that the immediate target of component technologies, perhaps technologies way down in the hierarchy (e.g., in the case of the computer technologies assembled in my laptop computer, technologies such memory circuits, screen drivers, compilers, parsing algorithms, etc.), is not the practice of the so-called ‘end user’ but the practice of the engineers who design and build specific laptop products (as configurations of component technologies). The point of these brief remarks on the systemic nature of technology is that, what we, in the context of CSCW, focus on are computing technologies at the ‘level’ of cooperative work practices. This does not just apply to the history of technology but equally to our understanding of the processes of technological development in which CSCW is engaged. Chapter 11: Formation and fragmentation 299 2. Computing technologies and cooperative work ‘By thinking about the history of “technology-in-use” a radically different picture of technology, and indeed of invention and innovation, becomes possible. A whole invisible world of technologies appears.’ Edgerton (2006, p. xi). From its very beginning, digital computing technology has been applied to cooperative work. Or rather, computing technology was not simply applied to cooperative work; the effort to meet various challenges of cooperative work has had significant impact on the development of the concept of computing. Addressing challenges of cooperative work has played a central role in developing ‘the computer’ into what we know today. I will briefly recount the major technological paradigm shifts in ordinary work settings. The point of departure is the development, prior to the industrial revolution, of systematic cooperative forms of work based on advanced division of labor that, in turn, was premised on having developed and mastered techniques of decomposition and recomposition of work processes However, this form of cooperative work had quite limited development potentials in terms of productivity, labor shortages, etc., and the development of machine technology can be seen as a means to overcome these restraints. And again, while the emergence of machinery at the point of production in the form of ‘machine tools’ and eventually ‘machine systems’ and concomitant forms of cooperative work relations showed enormous development potential, it too, eventually, had limitations rooted in the costs of constructing and modifying mechanical means of operational control. For that reason, mechanical machine technologies were, largely, confined to areas of mass production where the construction costs could be offset by high volumes. However, computing technologies, which were initially developed as a means of overcoming the bottleneck of increasingly complex cooperative forms of calculation work for scientific and engineering purposes, provided the technical basis for constructing and modifying control mechanisms at costs that were extremely low and that have, in fact, now been continually falling for half a century. The costs of constructing machine systems have in fact become so insignificant that relatively flexible machine systems for vast and varying arrangements of cooperative work have become economically viable. This is where the CSCW research program arises, not just as an academic pastime, but as a practical problem for modern industrial civilization. 2.1. Division of labor: Progressive forms of work organization The development of computing can only be understood when seen as an integral aspect of the development of work practices and their technologies over the last three centuries. 300 Cooperative Work and Coordinative Practices To make this claim understandable it is useful to briefly introduce the taxonomy of work organization that has been in widespread use in economic and technological historiography: (i) ‘craft work’ or domestic handicraft, (ii) the ‘putting out’ system or the ‘Verlagssystem’, (iii) the ‘manufactures’, and (iv) the machinebased ‘factory’ (e.g., Sombart, 1916; Braudel, 1979). The key distinguishing feature among these forms is the different arrangements of division of labor; although the machine-based ‘factory’ should be seen as a form of work organization in which the systematic division of labor in production is transcended, if only tentatively. A note of caution, before we move on. These progressive forms of work organization should not be thought of as stages, each stage leading inexorably to the next as a necessary step in the development of the work organization. The picture is far more complex than that. Firstly, technological development is ‘uneven’: different forms of work organization coexist across branches of production, across geographic regions, even within the production of the same product. Work organizations based on handicraft work continues in small scale production and is indeed reinvented time and again when novel kinds of products are first produced. One can for example observe the same forms at play in the history of the electronic and computer industries. In fact, in the modern automotive industry some components may be produced in domestic handicraft settings, others by advanced machinery, while the ultimate product is assembled in a work organization based on systematic specialization of handicraft. Secondly, retrograde development may take place, so that technologically backward forms become competitive again and gain a new lease of life when, for instance, large reservoirs of disposed peasants or underemployed workers in rural areas are created or become available or when dramatic technological changes put working conditions under pressure. That is, to repeat, the forms of work organization that we observe are not stages. Nor are they to be thought of as ‘models’ or ‘ideal types’. They are rather like recurring themes that are played again and again but in varying ways and contexts, often in the same sequence but sometimes not. So, instead of conceiving of these forms as stages, one might use a term such as progressive forms of work organization, if we let the term ‘progressive forms’ refer to two important features: On one hand, they represent a (precariously) cumulative process of development of organizational and technological knowledge. On the other hand, this cumulative effect is, of course, strengthened by the totalizing effect of the market. In so far as practical applications of this knowledge turns out to be successful in some recognized way (profitable, dependable, viable), the specific form of work organization feeds back as a competitive challenge to other actors in the market. Handicraft. In craft work, each worker typically masters the entire range of tasks involved in fabricating the ultimate product. In its simplest and most widespread form, craft work was an indispensable aspect of the subsistence economy of traditional rural life (e.g., spinning and weaving). These activities were performed by peasants as a sideline, that is, in periods or at times when agricultural tasks did not fully occupy their capacity. The unit of production was the individu- Chapter 11: Formation and fragmentation 301 al peasant household, with division of labor, if any, based on age and gender. In each of these ‘monocellular’ units, tasks were ‘undifferentiated and continuous’ (Braudel, 1979, p. 298). On the other hand, in domains of work that demanded a degree of skill that in turn required full-time engagement with the work (e.g., blacksmith, cutler, nailmaker, shoemaker, etc.), the work would not be carried out as a secondary occupation; specialization would be required. However, this type of craft work would be organized in a way rather similar to the sideline work in peasant households: as tiny family workshops, each with a master tradesman, twothree journeymen, and one or two apprentices. In both types, adult workers would possess the skills to undertake the entire range of tasks, and any division of labor between them would be occasioned and transient. The ‘putting out’ system. As an integral part of the development of the world market, especially from the 16th century, this ‘horde of little establishments, where family craft working was carried on’ (Braudel, 1979, p. 298), gradually became geared to the market and subsumed under new and systematic schemes of division of labor. Instead of being autonomous units working in parallel without any overriding plan or scheme, the individual units now became units of a larger scheme devised and coordinated by ‘merchant entrepreneurs’: ‘a number of individual units spread over a wide area but interconnected’ (ibid., p. 300). Coordinating the work of the many interconnected units, the merchant entrepreneur would advance the raw materials, take care of transportation of goods from one unit to the next (spinner, weaver fuller, dyer, shearer). ‘The pattern in every case was a sequence of manufacturing operations, culminating in the finished product and its marketing.’ (ibid.). This system for which no generally accepted term exists,84 developed from the end of the middle ages. It spread across the non-agricultural economy at large but became particularly dominant in the textile trades (spinning, weaving, etc.). However, the reason for paying attention to it here is its historical outcome; namely, that it promoted specialization at the level of production units. This was most pronounced in branches of production that fabricated ‘small and delicate objects’ such as lace and pins and needles, etc. (needles would for instance pass through 72 hands during the production process). But perhaps more importantly, the ‘putting out’ system occasioned merchant entrepreneurs and their assistants to develop practices of analyzing and coordinating work processes (standardization of materials, relative proportions of component processes, logistics) (Sombart, 1916). Manufactures: From the middle of the 16th century and towards the end of the 18th century the process of deepening and systematic division of labor took a new form: ‘manufactures’. Its characteristic feature was the centralization under one roof of the workforce. This made possible not only systematic supervision (and reduced costs of transportation etc.), but also ‘an advanced division of labor’ (Braudel, 1979, p. 300). What defines this movement is the systematic decompo84 The ‘putting-out’ system is sometimes dubbed ‘proto-industrialization’ in modern economic historiography (e.g., Mendels, 1972) but this notion is typically used as a designation for a stage in economic history (as pointed out by Coleman, 1984). 302 Cooperative Work and Coordinative Practices sition of received work processes into component processes and the corresponding specialization of individual workers to perform specific component processes. In many domains of work the form of work organization that characterized handicraft work, namely, that the individual worker mastered the entire range of tasks and that the different workers at the workshop could replace each other in performing different kinds of task, was transformed into an organization based on systematic division of labor in the workshop. In 1597 Thomas Deloney published a eulogy to the ‘the famous Cloth Workers in England’ as represented by the cloth maker John Winchcombe, ‘a famous and worthy man’. The text is of interest because it, as pointed out by the editor of Deloney’s works, shows ‘a detailed knowledge of Newbury, its surroundings, and the county families of Elizabethan Berkshire, which could only have been obtained by an actual residence there’ and, more to the point, because it offers a contemporary, if colorful and apologetic, description of the work organization in an early manufacture: ‘Within one roome being large and long, There stood two hundred Loomes full strong: Two hundred men the truth is so, Wrought in these Loomes all in a row. By euery one a pretty boy, Sate making quils with mickle ioy; And in another place hard by, An hundred women merrily, Were carding hard with ioyfull cheere, Who singing sate with voices cleere.’ (Deloney, 1597, pp. 20 f.). The poet goes on to enumerate the various handicrafts at this site by taking the reader through a imaginary tour of the establishment, pointing out — in addition to the 200 weavers, each assisted by a ‘pretty boy’, and the 100 joyfully singing carders —200 ‘pretty maids’ ceaselessly engaged in spinning in an adjoining chamber, and in another room 150 ‘children of poore silly men’ ‘picking wool’, and further, in yet other rooms, 50 male shearers, 80 rowers, 40 dyers, and 20 fullers. The manufactures were not only a form of organization that was seen as profitable and laudable; they were the object of scholarly interest, as a technology. The principles of systematic division of labor was formulated repeatedly the course of the 17th and 18th century, for instance by William Petty: ‘the Gain which is made by Manufactures, will be greater, as the Manufacture it self is greater and better. For in so vast a City Manufactures will beget one another, and each Manufacture will be divided into as many parts as possible, whereby the work of each Artisan will be simple and easie; As for Example. In the making of a Watch, if one Man shall make the Wheels, another the Spring, another shall Engrave the Dial-Plate, and another shall make the Cases, then the Watch will be better and cheaper, than if the whole work be put upon any one Man.’ (Petty, 1683, p. 472). Diderot went further and formulated the principles of the systematic division of labor in manufactures in his article on ‘Art’ in the first volume of the Encyclopédie. Discussing the advantages of manufactures, Diderot argued: Chapter 11: Formation and fragmentation 303 ‘The speed of work and the perfection of the product both depend entirely on the number of workers assembled. When a manufacture is large, each operation is dedicate to a different man. This worker here only makes one unique thing his entire life, that one there another thing, so that each operation is performed well and promptly, and so that each product, while the best made, is also the cheapest’ (Diderot, 1751, pp. 275 f.). However, the classic example of this form of work organization is undoubtedly that of pin manufactures, as given by Adam Smith in the first chapter of his Wealth of Nations. Arguing that the manufacture of pins, though ‘a very trifling manufacture’, is one ‘in which the division of labour is very often taken notice of’ and thus a paradigmatic case of ‘division of labour’, Smith goes on to describe the organization of a particular workshop: ‘One man draws out the wire, another straights it, a third cuts it, a fourth points it, a fifth grinds it at the top for receiving the head; to make the head requires two or three distinct operations; to put it on, is a peculiar business, to whiten the pins is another; it is even a trade by itself to put them into the paper; and the important business of making a pin is, in this manner, divided into about eighteen distinct operations, which, in some manufactories, are all performed by distinct hands, though in others the same man will sometimes perform two or three of them. I have seen a small manufactory of this kind where ten men only were employed, and where some of them consequently performed two or three distinct operations.’ (A. Smith, 1776b, pp. 4 f.). In reading Smith’s analysis it is important to realize that what he conceptualizes as division of labor is also, by the same token, cooperative work. The systematic division of labor is necessarily complemented by systematic cooperative work. This insight, which was implicit in Smith’s argument, was explicitly made by his followers such as Edward Wakefield and John Stuart Mill, although their discussion is confused because they, as dedicated followers of Smith, confounds the systematic and regular division of labor at the point of production with the disorganized and spontaneous division of labor at the level of the economy at large (cf., e.g., Mill, 1848, § I.VIII.1). The point is made more clearly by Mill’s contemporary, the German political economist Friedrich List who, as a protectionist, was prone to be critical of Smith: ‘a division of the operations, without the association of productive powers for a common end, would be of very little help in the production. That such a result may be obtained, it is necessary that the different individuals be associated intellectually and bodily and that they cooperate. He who makes the head of pins must count upon the labors of him who makes the points, so that he does not run the risk of making pin heads in vain. The work activities should be in a suitable proportion to each other; the workers ought to be located as near to each other as possible; and their cooperation should be insured.’ (List, 1841a, p. 165; cf., List, 1841b, p. 231). For our purposes, this insight is important. Cooperative work is found in the entire course of human history, but until the systematic division of labor in manufactures it either occurred sporadically (e.g., hunting of large game, clearing of forest) or was marginal to ordinary production (construction of sacral buildings, fortifications, dams, roads), and the work organization of the manufactures is thus the first example systematic and continual cooperative work at the point of production. Smith was of course not particularly interested in the technological and managerial issues of work organization, but in the effects of work organization on the 304 Cooperative Work and Coordinative Practices creation of ‘wealth’, and he therefore quickly moved on to expound three advantageous effects of division of labor (A. Smith, 1776b, p. 7): ‘First, the improvement of the dexterity of the workman necessarily increases the quantity of the work he can perform; and the division of labour, by reducing every man’s business to some one simple operation, and by making this operation the sole employment of his life, necessarily increased very much dexterity of the workman.’ However, in his analysis Smith stayed well within the limits of the manufactures, emphasizing the advantages over artisanal work but blind to the inherent limitations of the manufactures. He made some very enthusiastic claims concerning the gains in productivity as a result of division of labor. To determine the effect of division of labor on productivity, Smith compared his case with that of a single polyvalent worker: ‘Those ten persons […] could make among them upwards of forty−eight thousand pins in a day. Each person, therefore, making a tenth part of forty−eight thousand pins, might be considered as making four thousand eight hundred pins in a day. But if they had all wrought separately and independently, and without any of them having been educated to this peculiar business, they certainly could not each of them have made twenty, perhaps not one pin in a day; that is, certainly, not the two hundred and fortieth, perhaps not the four thousand eight hundredth part of what they are at present capable of performing, in consequence of a proper division and combination of their different operations.’ (A. Smith, 1776b, p. 7). That is, Smith asserted that the production of pins, under the scheme of division of labor, had increased labor productivity by a factor 4,800, or at least 240. However, the claim that a single worker, on his own, would hardly be able to produce one pin per day, leave alone twenty, was sheer guesswork on Smith’s part. In fact, in artisanal pin-making in France of the 18th century an apprentice should be able to produce 1,000 pins in half a day, or 2,000 on a daily basis (Peaucelle, 2007, p. 196). A 100 percent increase in productivity is of course significant by any standard, but it is not the mind-blowing rate that has mesmerized Smith’s readers since then. The point of this is not to be pedantic but rather that the manufactures work organization was severely limited by the characteristics of the human sensorymotor system. Productivity may increase as a result of the increase in dexterity of the individual but with a few weeks or of training, the learning curve would reach a plateau. As a second advantage of division of labor, Smith pointed to the possibility of ‘saving the time commonly lost in passing from one sort of work to another is much greater than we should at first view be apt to imagine it’. Productivity will of course also increase as continuity of operation eliminates the cost of transiting from one operation to the next (set-up costs, in modern terminology) but, again, when the set-up costs have been eliminated the productivity rate will also flatten out. And as the third and last advantage, Smith argued that specialization engenders technical innovation: ‘everybody must be sensible how much labour is facilitated and abridged by the application of proper machinery. It is unnecessary to give any example. I shall only observe, therefore, that the Chapter 11: Formation and fragmentation 305 invention of all those machines by which labour is so much facilitated and abridged seems to have been originally owing to the division of labour. Men are much more likely to discover easier and readier methods of attaining any object when the whole attention of their minds is directed towards that single object than when it is dissipated among a great variety of things.’’ (A. Smith, 1776b, pp. 7-10.). Again, despite the use of the term ‘machinery’, Smith’s analysis remained firmly entrenched within the conceptual horizon of the manufactures. When talking about ‘machinery’ he was not talking about what we today understand by machinery, namely devices that can operate automatically (under human supervision, of course, but without continual human involvement in the transformation process). This issue will be discussed at length later, but it has to be pointed out here that Smith, like his contemporaries, was using the terms ‘machine’ and ‘machinery’ very loosely to denote any kind of composite tool that somehow augments human capacity such as, for example, ‘the ship of the sailor, the mill of the fuller, or even the loom of the weaver’ (A. Smith, 1776b, p. 12). The way in Smith presents the case (‘I have seen a small manufactory of this kind’) and not least the way he has been read may lead one to assume that he in his analysis described something that to his contemporaries was breaking news. It was not. As pointed out above, the manufactures form of work organization had been the object of study for a century before Smith described the pin manufacture. More than that. The pin manufacture had been a case of special interest to scholars several decades before Smith wrote about it. And although Smith claimed to have observed the workshop he described, the evidence indicates that this is actually not true (Peaucelle, 2006, 2007). As Smith himself pointed out, the work organization of pin making had, at the time, ‘very often’ been ‘taken notice of’. In fact, the division of labor in the manufacture of pins had been described long before Smith made use of the case (in his Lectures on Jurisprudence from 1762-64 and in his Wealth of Nations) and it was common knowledge among scholars who had an interest in technology and economy. Thus, Ephraim Chambers’s Cyclopædia from 1728, well-known and respected at the time, emphasized the advanced division of labor in the manufacture of pins.85 The primary sources for Smith’s description were French. As mentioned earlier, the French Académie des science had since 1675 been collecting data concerning manufacturing processes based on field reports from local observers who were attached to the prefectures but were doing their research at the direction of scholars attached to the L’Académie (Peaucelle and Manin, 2006). One of the production processes on which reports were collected (since 1693) was that of the pin manufactures in Normandy (Peaucelle and Manin, 2006; Peaucelle, 2007). The analyses were eventually made publicly available in a various documents, first of all in one of the volumes of Descriptions des arts et métiers, a collection of writings devoted to the Art of pin-making edited by Duhamel de Monceau and containing texts by 85 ‘Notwithstanding that there is scarce any Commodity cheaper than Pins, there is none that passes thro’ more hands e’re they come to be sold. — They reckon 25 Workmen successively employ’d in each Pin, between the drawing of the Brass-Wiar, and the sticking of the Pin in the Paper’ (Chambers, 1728). 306 Cooperative Work and Coordinative Practices Réaumur, Perronet, and Chalouzières (Réaumur, et al., 1761) but also, and better known, in the very detailed accounts in Diderot’s Encyclopédie (Delayre, 1755; Perronet, 1765). The findings of these reports and analyses were then further distributed in various forms: in journal reviews (e.g., in the Journal des sçavans, 1761), in pocket dictionaries (such as Macquer’s Dictionnaire portatif des Arts et Métier, 1766). In short, the work organization of manufactures was the topic of great interest, and the organization of the manufacture of pins was becoming a paradigm case for the technological and managerial knowledge of this kind of work organization. Chapter 11: Formation and fragmentation 307 Figure 1. Illustrations of the pin making manufacture from Diderot’s Encyclopédie (the supplementary series of plates: Recueil de planches, sur les sciences, les arts liberaux, et les arts méchaniques, avec leur explication, vol. 3, Paris, 1765). Pins were made from brass wire that is drawn and cut and so forth. The topmost plate shows winding, unwinding, and washing coils of wire before drawing it. The second plate shows the processes of drawing the wire (right), cutting it (center), and pointing the pins (left). The bottom plate depicts the final process of heading the pins by annealing (right, at the back) and hammering (at the contraption in the center). The pins are finally tin-coated, washed, polished, and packed (left). The pin making process was described in detail by Delayre in vol. 5 of l’Encyclopédie (1755), by Duhamel du Monceau in a volume of the Descriptions specifically devoted to the art of making pins (Réaumur, et al., 1761), and by Perronet in vol. 3 the Recueil de planches (1765). (Cf. also Gillispie, 1959, vol. 1, plates 184-186 and associated text). 308 Cooperative Work and Coordinative Practices That is, the form of cooperative work that is characteristic of the manufactures was investigated, analyzed, and referred to by contemporaries as extant and ongoing cases; they became an element of common knowledge: hailed by poets and propagandists such as Deloney, studied and described in detail by scholars such as Delayre and Perronet, and ultimately elevated to paradigm status by Adam Smith. The manufactures play an important historical role in the development of machinery in production processes. The analysis of work processes involved in the decomposition of craft work, the standardization of methods for each constituent process, and their recomposition and planned integration paved the way for the mechanization of constituent processes. Automatic machines had, of course, existed for centuries at this time. Clocks are a case in point. At the point of production, however, machinery only became significant by the end of the 18th century. Now, manufactures were never simply transformed into machine-based factories, nor did machinery always emerge from manufactures. Mechanization could also occur from the construction of machinery to automate an entire process. However, the manufactures play a critical role in the development of machinery for the automatic enactment of production processes, in that it developed widespread and (fairly) systematic knowledge of the principles of division of labor: The concept of division of labor, of sequences of elementary operations, of productivity as measured by output per worker per day, of different cadencies in a composite process, of ‘line balancing’, etc. The transformation of craft work into a form based on systematic division of labor can be seen as having been instrumental in developing the requisite technical and managerial practices for mechanization of production to be conceivable as a practical option. That is, the very idea that a received work practice could be analyzed, decomposed, and composed anew was the decisive concept. In fact, our very concept of technology is predicated on this insight. In the development of technology — of technical knowledge, that is — Smith’s version of the story is hugely important, since it provided the paradigmatic account: that a concerted arrangement of workers, each specialized in performing a simple operation in a systematically planned sequence, could achieve a productivity greatly surpassing that of the same number of skilled workers, even artisans, working in parallel. More importantly, it highlighted and directed the attention of engineers, technologists, and managers as well as scholars to the fact that in a carefully planned division of labor, the collaborating workers, as a concerted collective, could master a process that none of them could master and perhaps even account for individually. This was underscored, albeit indirectly, by Charles Babbage when he, in a comment on Adam Smith’s analysis of the advantages of the division of labor in manufactures, made some critical corrections. Going back to one of sources used by Smith (namely the abovementioned Perronet), Babbage points out that a major factor making division of labor economically advantageous is the circumstance that, whereas workers with the skills of artisans, and corresponding salaries, would be required for the most complex operations, operations, Chapter 11: Formation and fragmentation 309 such as heading pins, could be performed by workers with very little training and similar wages (women and children): ‘it appears to me, that any explanation of the cheapness of manufactured articles, as consequent upon the division of labour, would be incomplete if the following principle were omitted to be stated. That the master manufacturer, by dividing the work to be executed into different processes, each requiring different degrees of skill or of force, can purchase exactly that precise quantity of both which is necessary for each process; whereas, if the whole work were executed by one workman, that person must possess sufficient skill to perform the most difficult, and sufficient strength to execute the most laborious, of the operations into which the art is divided.’ (Babbage, 1832, § 226). This has been much discussed, and rightly so, for it is an important corrective to Smith’s analysis of division of labor. The point in our context, however, is that Babbage here underscores a lesson that is easily ignored: that the individual worker in such arrangements only needs to understand and master his or her partial task and that the partial task might be made so simple that it can be mechanized. From this insight emerges the modern concept of the control function, predicated on a distinction of planning and execution of operations. The systematic division of labor that characterizes manufactures has inherent limitations, as pointed out by Marx: ‘For a proper understanding of the division of labour in manufacture, it is essential that the following points be firmly grasped: First, the analysis of the production process into its specific phases coincides, here, strictly with the dissolution of a handicraft into its various partial operations. Whether complex or simple, the performance retains the character of handicraft and, hence, remains dependent on the strength, skill, quickness, and sureness, of the part-worker in handling his instruments. The handicraft remains the basis. This narrow technological basis excludes a really scientific analysis of the production process, since each partial process to which the product is subjected must be capable of being carried out as partial handicraft. Precisely because handicraft skill, in this way, remains the foundation of the process of production, each worker becomes exclusively appropriated to a partial function and that, for the duration of his life, his labour power is turned into the organ of this detail function.’ (Marx, 1867a, pp. 274 f. ).86 The manufactures work organization was not only severely limited by the characteristics of the human sensory-motor system and that the learning curve would reach a plateau; it was also, Marx argued (with Ure), limited by its severe dependence on the handicraft worker and thus that worker’s resistance to attempt to speed up the process. Still, the fact that the performance retains the character of handicraft has not prevented foremen and engineers in the twentieth century from taking specialization to even greater heights than that represented by the classic manufactures. In order to fully exhaust the performance potential of workers, work processes and operations were subjected to careful systematic analysis and redesign, and for this 86 Author’s translation from the German original. The standard English translation (by Moore and Aveling) is unfortunately rather inaccurate and, indeed, occasionally misleading. Anyway, the English translation of the relevant passages can be found in Chapter XIII (‘Co-operation’), Chapter XIV (‘Division of labour and manufacture’), and Chapter XV (‘Machinery and modern industry’), esp. Section 1 (‘The development of machinery’, pp. 374 ff.). (Marx, 1867c). 310 Cooperative Work and Coordinative Practices purpose engineers such as Frederick W. Taylor (1911) Henri Fayol (1918) and their many followers (e.g., Urwick and Brech, 1945, 1946, 1948) developed and applied advanced methods and techniques of observation such as stopwatches, video cameras, etc., as well as methods of work design such as instruction cards and slide rules, etc. Likewise, to optimize the performance of the human sensorymotor system physiologists became involved in the design of elementary operations, tools, and work stations and eventually developed ‘human factors’ or ergonomics as a engineering discipline with its own repertoire of methods and techniques. In a parallel attempt to overcome workers’ resistance, social psychologists became involved in developing methods of increasing workers’ ‘motivation’. This effort is famously represented by the so-called ‘Hawthorne experiments’ carried out between 1924 and 1933 (e.g., Roethlisberger and Dickson, 1939). Where Taylor consistently argued for financial means for motivating workers, the social psychologists concluded from these ‘experiments’ that measures such as friendly supervision and ‘social relations’ in ‘work groups’ were of overwhelming importance. However, more recent and more rigorous studies comparing the Hawthorne conclusions and the Hawthorne evidence show ‘these conclusions to be almost wholly-unsupported’ (Carey, 1967) and that there is ‘no evidence of Hawthorne effects’ (Jones, 1992). This has not prevented the formation of a ‘human relations’ profession with its own repertoire of methods and tests, however (for discussions of this phenomenon of managerial ideology, cf. Gillespie, 1991; Brannigan, 2004). While initially focused on increasing daily output per worker by optimizing the part-performance of the part-worker, the methods and techniques developed by the ‘scientific management’ school were not restricted to that. Major improvements in overall productivity have been obtained by optimizing the flow of work pieces from one part-worker to the next, in particular by motorizing the transportation of work pieces, typically by means of conveyor belts. The assembly line one finds in, e.g., automobile assembly plants, and which many sociologist of work have taken as exemplary of the machine-based factory system, is in fact and on the contrary the most advanced form of the systematic division of labor exemplified by the pin manufacture. The confusion is nurtured by imagery in which workers, to the naïve eye, appear as small cogwheel in a vast automaton and it is reinforced by overindulgence in metaphorical discourse where certain organizational forms are talked of a ‘machines’.87 But even superficial observation of the organization of work in these plants shows that is generally that of extreme division of labor (for a collection of historical photos, cf. Marius Hammer, 1959). What became ‘mechanized’ in automobile assembly and in other kinds of assembly work in the course of the 20th century was the transportation of parts and assemblies between workstations, not the operations at the various work stations, 87 The distinguished historian Siegfried Giedion, to take but one example, posited that ‘The symptom of full mechanization is the assembly line, wherein the entire factory is consolidated into a synchronous organism’ (1948, p. 5). Chapter 11: Formation and fragmentation 311 and the ‘mechanization’ of the transportation system merely consisted in a conveyer belt arrangement powered by electrical motors. In fact, it is only quite recently, after the microprocessor technology began to stabilize in the late 1970s, that self-acting machinery is being applied in significant ways in the automotive industry (e.g., welding robots). Moreover, the assembly line was never representative of industrial production, not even of mass production. Because the assembly line requires a large investments in an inflexible workflow layout, it was and remains an exceptional case in modern manufacturing. In the words of Taylor’s biographer, Robert Kanigel: ‘While the assembly line remains a common if tired metaphor, it defines surprisingly little of modern manufacturing. The Taylorized workplace, on the contrary, appears everywhere, heedless of the lines between one industry and the next’. In short, ‘Fordism was the special case, Taylorism the universal’ (1997, p. 498). Later in the twentieth century, engineers such as Taiichi Ohno of Toyota developed and refined the repertoire of methods and techniques developed by the ‘scientific management’ school (Ohno, 1988). Although the principles he and his colleagues developed, widely known as the ‘Toyota production system’, are hailed as a break with ‘Taylorism’, they are at least as rigorous in their attention to systematic analysis of operations and continuity of the flow of work as previous forms of ‘scientific management’. What is new is rather, on one hand, the refinement of the manufactures form of work organization to take into account other performance criteria than output per worker per day, namely, criteria such as the flexibility and responsiveness of the overall cooperative work effort (by ‘just-intime’ principles) as well as product quality (cf. Peaucelle, 2000), and on the other hand a strong emphasis on ‘continual improvement’ and the involvement of workers in achieving this (cf. Spear and Bowen, 1999), Marx’s observation remains valid: the performance of cooperative work based on advanced de- and recomposition of labor ‘remains dependent on the strength, skill, quickness, and sureness, of the part-worker in handling his instruments’. With the introduction of machinery, this dependence is broken and with this also the dependence of human productivity on progressive deepening of the division of labor. 2.2. Machinery: The issue of the control function The factory, the form of organization of work based on mechanization of the constituent work processes, is typically seen as yet another progressive form of work organization. The reason for this is that the advanced division of labor, based on decomposition of received handicraft and the recomposition of the production process as a systematically designed process, historically provided the technology of work analysis that is an essential foundation of machinery. At the same time, however, the factory represents a form of work organization that cannot be grasped as yet another permutation of handicraft and division of labor. 312 Cooperative Work and Coordinative Practices Mechanization can be conceived of as a process that unfolds in two dimensions: on one hand the process of transfer of operational control from workers to (thereby increasingly ‘self-regulating’) technical implements, i.e., machines, and on the other hand the process of technical integration of multiple machines into machine systems. It is the latter process, the development of machine systems, that is of particular concern here, as computational coordination technology is a special but now dominant type of machine system. In manual work, as exemplified by traditional craft work, the process of transformation —the operation of the grindstone, spinning wheel, handloom — is literally ‘in the hands’ of the worker. Some external source of power may be employed (e.g., animal, wind, water, steam), but the operation of the tool (stone, spindle, shuttle) is controlled by the worker, and the performance thus depends on his or her skills. With machinery, the control of the operation of the tool is performed ‘automatically’, by the implement, without human continual intervention or mediation. This conception is a modern one; it originates in the industrial revolution. One of the first to make this distinction was Charles Babbage who, for most of the 1820s, conducted extensive field work in ‘a considerable number of workshops and factories, both in England and on the Continent, for the purpose of endeavouring to make myself acquainted with the various resources of mechanical art’ (Babbage, 1832, p. iii). Faced with ‘the wide variety of facts’ obtained at these visits, he found it ‘impossible not to trace or to imagine’ ‘some principles which seemed to pervade many establishments’, in particular the principles of division of labor and the ‘mechanical principles which regulate the application of machinery to arts and manufactures’ (ibid., pp. iii f.). It was, presumably, his extensive exposure to the facts of the ground, coupled with his skilled ability for generalization, that then enabled him to see, with some clarity, what was specific in the new machine technologies that were then being developed and applied, namely that a set of tools are ‘placed in a frame’, moved ‘regularly by some mechanical contrivance’, and thus ‘acted on by a moving power’ (ibid., §§ 10, 224, pp. 12, 174): ‘When each process has been reduced to the use of some simple tool, the union of all these tools, actuated by one moving power, constitutes a machine.’ (ibid, § 225, p. 174). However, the first to express the notion of ‘control’ succinctly was Andrew Ure in 1835. Like Babbage before him, he had engaged in what we today would call field studies in the factory districts of England. In his own words, he ‘spent several months in wandering through the factory districts of Lancashire, Cheshire, Derbyshire, &c., with the happiest results to his health; having everywhere experienced the utmost kindness and liberality from the mill-proprietors’ (Ure, 1835, p. ix). Having returned from his field trip, he summarized his findings in a book that, while shamelessly apologetic of the factory regime, contains astute observations of the then emerging technologies of ‘automatic’ production: ‘The principle of the factory system […] is, to substitute mechanical science for hand skill, and the partition of a process into its essential constituents, for the division or graduation of labour among Chapter 11: Formation and fragmentation 313 artisans. On the handicraft plan, labour more or less skilled, was usually the most expensive element of production […] but on the automatic plan, skilled labour gets progressively superseded, and will, eventually, be replaced by mere overlookers of machines.’ (Ure, 1835, p. 20) The concept underlying Babbage and Ure’s distinction is that of what we today call the control function and the related notion of transfer of control from human to implement; that is, the (partial) elimination of human labor in so far as the actual transformation process is concerned. Simply put, on this view a machine differs from a tool by being able to perform autonomously, within certain limits, by virtue of a mechanical control function.88 Following Babbage and Ure, but conceptually more stringent, Marx made the concept of the control function the cornerstone of his analysis of the transformation of work in the course of the industrial revolution. The classic statement on machinery is in the discussion of ‘Machinery and big industry’ in Capital: ‘On a closer examination of the machine tool or the working machine proper, one recognizes by and large — though often, no doubt, in very modified form — the apparatus and tools used by the handicraftsman or the worker of manufactures, but instead of as tools of humans, now as tools of a mechanism, or mechanical tools. […] The working machine is therefore a mechanism that, after being set in motion, performs with its tools the same operations that were formerly done by the workman with similar tools. Whether the motive power is derived from humans or from some other machine, makes no difference in this respect. After the transfer of the tool proper from humans to a mechanism, a machine replaces a mere tool. The difference strikes one at once, even when the human remains the prime mover.’ (Marx, 1867a, pp. 303 f.). This was not an casual remark. Marx had first read Ure and Babbage forty years earlier (Marx, 1845b, pp. 325-351) and had made this concept of machinery a pivotal element of his understanding of contemporary developments. Shortly after these initial studies he wrote a small book in which he, quoting Babbage and Ure, stated rather unequivocally that ‘What characterizes the division of labor in the automatic workshop is that labor has there completely lost its character of specialism. But the moment every special development stops, the need for universality, the tendency towards an integral development of the individual begins to be felt. The automatic workshop wipes out specialists and craft-idiocy.’ (Marx, 1847, chapter 2.2). And again, ten years later, in the first outline of his economic theory, the Grundrisse from 1857-58, he makes this point forcefully: ‘Work no longer appears so much as included within the production process; but rather the human engages in that process as its overseer and regulator. […] It is no longer the worker who interposes the modified natural object [i.e., the tool] as an intermediate between the object and himself; but rather, he now interposes the natural process, which he transforms into an industrial one, as a means between himself and inorganic nature, which he masters. He stands beside the production process, rather than being its main agent.’ (Marx, 1857-58a, p. 581). He subsequently reread Babbage and Ure in the context of the extensive studies of the development of production technologies and work organization (cf. Marx, 88 Charles Kelley (1968) offers a good introduction to the modern concept of control. 314 Cooperative Work and Coordinative Practices 1861-63, pp. 229-318, 1895-2090) that found their ultimate expression in the chapters on ‘Cooperative work’, ‘Division of labor’, and ‘Machinery’ in Capital. In view of the fact that the mathematical theory of control functions had not been proposed yet when Babbage, Ure, and Marx offered these initial formulations of the modern concept of machinery, this is of course quite remarkable. Their ability to do so, however, was not the result of prescience but was a consequence of their keen interest in what we today would call the changing allocation of function between human and machine which they considered of the utmost economic and social importance. This methodological preferences is made explicit at several places, for instance in a letter from 1863 in which Marx summarized his initial conclusions from a thorough study the development the progressive forms work organization and especially machinery: ‘there is considerable controversy as to what distinguishes a machine from a tool. After its own crude fashion, English (mathematical) mechanics calls a tool a simple machine and a machine a complicated tool. English technologists, however, who take rather more account of economics, distinguish the two (and so, accordingly, do many, if not most, English economists) in as much as in one case the motive power emanates from man, in the other from a natural force. […] However, if we take a look at the machine in its elementary form, there can be no doubt that the industrial revolution originates, not from motive power, but from that part of machinery called the working machine by the English, i.e., not from, say, the use of water or steam in place of the foot to move the spinning wheel, but from the transformation of the actual spinning process itself, and the displacement of that part of human labor that was not mere exertion of power (as in treading a wheel), but was concerned with processing, working directly on the material to be processed. […] To those who are merely mathematicians, these questions are of no moment, but they assume great importance when it comes to establishing a connection between human social relations and the development of these material modes of production.’ (Marx, 1863, pf. 320 f.). (Cf. also Marx, 1861-63, pp. 1915-1917). Now, what is interesting from a CSCW perspective, as opposed to a history of science and technology point-of-view, is not these early developments of the concept of machinery. What is very relevant to CSCW is rather the concept of machine systems these authors suggested. In formulations that are strikingly visionary, Ure observed: ‘The term Factory, in technology, designates the combined operation of many orders of workpeople, adult and young, in tending with assiduous skill a system of productive machines continuously impelled by a central power. This definition includes such organizations as cotton-mills, flax-mills, silk-mills, woollen-mills, and certain engineering works; but it excludes those in which the mechanisms do not form a connected series, nor are dependent on one prime mover. But I conceive that this title [the term Factory], in its strictest sense, involves the idea of a vast automaton, composed of various mechanical and intellectual organs, acting in uninterrupted concert for the production of a common object, all of them being subordinated to a self-regulated moving force.’ (Ure, 1835, pp. 13 f.). In his conception of machine systems Marx again followed Ure, but his conception of machine system is significantly more developed. This of course reflects the rapid development of machine technology had undergone since Ure was roaming the factory districts of England some 33 years earlier. Based on studies of ‘au- Chapter 11: Formation and fragmentation 315 tomatic workshops’ as represented by paper and envelope factories, power looms, and printing presses, Marx stated: ‘A proper machine system only takes the place of the particular independent machines, where the work piece [Arbeitsgegenstand] undergoes a continuous series of different process steps, carried out by a chain of machine tools that, while of different species, complement one another.’ (Marx, 1867a, p. 309) (cf. also Marx, 1861-63, pp. 1940-1946). He compared this system of interoperating machines with the division of labor that were characteristic of the manufactures: ‘The cooperation by division of labor that characterizes Manufacture here reappears, only now as a combination of partial working machines. The specialized tools of the various specialized workmen, such as those of the beaters, cambers, twisters, spinners, etc., in the woollen manufacture, are now changed into the tools of specialized machines, each machine constituting a special organ, with a particular function, in the combined tool-mechanism system.’ (Marx, 1867a, p. 309). However, Marx went on, ‘an essential difference at once manifests itself’: ‘In the manufacture, it must be possible for each particular partial process to be performed by workers, individually or in groups, by means of their manual tools. While the worker is adapted to the process, then on the other hand, the process was previously adapted to the worker. This subjective principle of division of labor vanishes in production by machinery. The process as a whole is here considered objectively, in and for itself, analyzed into its constituent phases; and the problem of how to execute each partial process and connect the different partial processes into a whole, is solved by technical application of mechanics, chemistry, etc.’ (Marx, 1867a, pp. 309 f.). According to Marx then, in the advanced factory, based on machine systems, the worker is no longer subsumed under the regime of extreme specialization that characterizes the manufactures but become a member of a cooperative ensemble that, as a collective, supervises the operation of the machine system: ‘The combined working machine, now an organized system of different species of individual machines and of groups of such machines, becomes increasingly perfect in so far as the process as a whole becomes a continuous one, i.e., the less the raw material is interrupted in its passage from its first phase to its last; in other words, the more its passage from one phase to another is conveyed by the mechanism itself, not by human hand.’ (Marx, 1867a, p. 310). Marx here clearly conceived of machine systems in terms of a system of automatically interacting machines, as a system of interoperating control functions. With the development of systems of machines that interoperate, cooperative work becomes (to some extent and at different levels of granularity) mediated and regulated by these machine systems: ‘The implements of labor acquire in machinery a material mode of existence that implies the substitution of human force by natural forces and of experience-based routine by the conscious application of science. In the manufacture, the organization of the social work process is purely subjective, a combination of partial workers; in the machine system, modern industry creates an objective productive organism, which the worker meets as an existing material condition of production. In simple cooperation, and even in cooperation predicated on division of labor, the displacement of the individualized worker by the socialized worker still appears to be more or less accidental. Machinery […] operates only in the hand of directly associated or communal work. Hence the cooperative character of the work process now becomes a technological necessity dictated by the nature of the implement of work itself.’ (Marx, 1867a, p. 315). 316 Cooperative Work and Coordinative Practices With the industrial mode of production, on this view, cooperative work is more than an economically advantageous arrangement; it is ‘a technological necessity’. The machine system presumes a cooperative work arrangement for its operation and the individual activities of the cooperating workers are in turn mediated and regulated by the machine system. Figure 2. Machine system: The printing press of The Times, c. 1851, built by Cowper & Applegath. From a ‘Survey of the Existing State of Arts, Machines, and Manufactures’ by a Committee of General Literature and Education appointed by the Society for Promoting Christian Knowledge (1855, p. 211), one of Marx’ primary sources on ‘machine systems’. While remarkably modern, the conception of machine systems developed by contemporary analysts such as Ure and Marx is, not surprisingly, limited by the practical horizon of the factory system of the 19th century. As is evident from the above quotations, Ure’s conception of ‘the factory’ is predicated on not only ‘the idea of a vast automaton, composed of various mechanical and intellectual organs, acting in uninterrupted concert for the production of a common object’ but also on the premise that ‘all of them being subordinated to a self-regulated moving force’. What, on Ure’s conception, makes the ‘system of productive machines’ a system, ‘a vast automaton’, is the fact that multiple machines are being ‘continuously impelled by a central power’ (Ure, 1835, pp. 13 f.). The same ambiguity can be found in Marx: ‘A system of machinery, whether based on mere cooperation of working machines of the same kind, as in weaving, or on a combination of machines of different kinds, as in spinning, constitutes in and for itself one huge automaton, as soon as it is driven by a self-acting prime mover. […] The machinery-based enterprise achieves its most developed form as an organized system of working machines that receives its motion from a central automaton by transmission machinery.’ (Marx, 1867a, pp. 310 f.) Chapter 11: Formation and fragmentation 317 The influence from Ure is obvious in that Marx, in discussing machine systems, also made a ‘prime mover’ the defining feature of machine systems, as opposed to control of operations. The picture of the central steam engine that, while entrenched in the basement of the factory, drives all machines in the factory via a highly visible and comprehensive power-transmission system of driving belts, shafts, and gear trains, surely must have been as evocative then as it is now. It is nevertheless, prima facie, somewhat puzzling that Marx, who confidently and clearly stated that ‘the industrial revolution originates, not from motive power, but from that part of machinery called the working machine’ and that a ‘the elimination of that part of human labor that was not mere exertion of power […], but was concerned with processing, working directly on the material to be processed’, would include ‘a self-acting prime mover’ as defining of a machine system. This ambiguity is not accidental, however. It is rooted in historically given conceptual limitations. The transmission of power to the tool and the automatic control of the movements of the tool had not yet been technically separated. Ure and Marx therefore basically defined ‘machinery’ in terms of the role of the worker vis-à-vis the implement as such, that is, in terms of the ‘self-acting’ character of its operation, not in terms of specific technical features. For Marx the defining feature was that it is the machinery, not the worker, that immediately controls the movements of the tool. The notion of a distinct control mechanism was absent from the reasoning of Marx. For good reasons. First of all, in spite of his attention to technical issues in the historical development of machinery, his primary concern was the changing role of workers in production, for, as he put it in the letter cited above, the questions of the role of the worker vis-à-vis the implement was ‘of no consequence’ to ‘pure mathematicians’, ‘but they become very important when it is a question of demonstrating the connection of human social relations to the development of these material modes of production’ (Marx, 1863, p. 321). Thus, when he, following Ure, made the ‘prime mover’ a defining feature of ‘machine systems’, he was considering the obvious fact that, while a worker might be the source of energy with a single working machine, as had been the case for centuries in the case of the spinning wheel, it was practically impossible for a worker to move a connected system of machines. That is, a continuous ‘external’ supply of energy to drive the machine system was simply a necessary precondition for these to be workable. Consequently, while his definition of machinery as such is clearly based on the notion of a control function, this concept is abandoned when the phenomenon of machine systems is considered. Distinct control mechanisms, physically separate from the energy transmission systems of the implement, were extremely rare. Generally the control function was performed by the physical union of the mechanism of transfer of power and the mechanism of controlling the movements of the tool. This practico-technical circumstance made it difficult to formulate and apply a strict concept of control, for Ure and Marx and for generations of technologists too. 318 Cooperative Work and Coordinative Practices For more than a century, machine systems were a rare phenomenon, restricted to a limited set of branches of mass production. The reason for this is of critical importance for understanding the role of computing technologies and, by implication, CSCW. In early machines, such as Richard Roberts’ ‘self-acting mule’ (1825-30), the control of the behavior of the tool was completely integrated with the system of transmission and transformation of energy. That is, the movement of the tool and the workpiece was regulated by the very same parts of the machine (driving belts, rack-and-pinion, gears, clutches, camshafts, crankshafts, etc.) that transferred energy to the tool (or the workpiece), that is, made it move. For reasons of economy and reliability, the overwhelming concern of mechanics would therefore be to keep the number of parts to a minimum. In the words of Larry Hirschhorn in his brilliant study of mechanization and computer control in the work place: ‘in a good mechanical design the same part or series of parts simultaneously transmits power, transforms motion, and controls the speed and direction of movement, in this way minimizing the number of parts and preventing unwanted action’ (Hirschhorn, 1984, p. 16). Because of this, it is difficult and expensive to modify the design of such machines. This is crucial for understanding the characteristics of mechanical technologies and, by contrast, of computing. As Hirschhorn puts it: ‘well-designed machines are also highly constrained ones, single-purpose in character and design and hard to modify. […] In general, since the systems of transmission, transformation, and control share the same parts, modifying one system inevitably means modifying the others. […] In becoming more productive, they lose flexibility’ (Hirschhorn, 1984, p. 18). As a result, for more than a century the domains in which machinery could be applied productively were quite limited. Machinery was generally only applied to branches of mass-production where the investment in special-purpose machinery could be amortized. And as far as machine systems are concerned, the scope of productive application was even more limited. The cost of building and modifying machine systems prevented machine system technology to spread beyond a few branches of mass-production such as, for example, the production of envelopes, newspaper printing, chemical production, power production and distribution, metallurgy and of course transportation. For machines to be effectively flexible requires the mechanism of control be ‘physically separate’ from the mechanism of energy transmission (O. Mayr, 1969). In retrospect it is possible to discern distinctive control mechanisms in the form of devices such as float valves (devised by, e.g., Ctesibios in the 3rd century BCE, Heron in the first century CE, Banū Mūsā in the 9th), thermostats (by e.g., Drebbel in the 17th century), regulators in windmills (by British millwrights in the 18th century), and speed regulators in steam engines (Watt) (O. Mayr, 1969). But it is significant that even Watt does not seem to have recognized the ‘universality’ of his invention. Not only did he not think of it as a ‘new invention’ but merely as an application of an invention made by others for the regulation of water and windmills. But he does not seem to have commented on the underlying ‘feedback’ principle. This leads the authority on the history of control mechanisms, Otto Chapter 11: Formation and fragmentation 319 Mayr, to conclude: ‘One might infer from his silence that he did not see anything particularly interesting in this principle. Compared with the large but straightforward task of producing power, the regulating devices and whatever theory they involved may have appeared to the sober Watt as secondary, if not marginal’ (O. Mayr, 1969, pp. 112 f.). The prospect of mastering a vast reservoir of motive power and to be able to apply it to augment human productive capacity was so dominant that it was difficult to discern that something ultimately stupendous was underway: the development of technologies of control systems. Even those who were understood the essentials of the paradigm shift — Ure and Marx — did not get it quite right. When discussing machine systems and what made the interconnected machines a system, they were still limited by the concept of the ‘prime mover’. The conceptual distinction between the mechanism of provision of energy and the mechanism of control, or between ‘energy flow’ and ‘control flow’, only became articulated and stable as the distinction was made in practice: when machine builders in actual practice began to physically separate the mechanism for provision of energy from the mechanism for control of operations. And in actual practice the distinction was only made in the course of a very long process of technological development.89 For a century only sporadic progress was made in terms of physically separate and hence easily replaceable control mechanisms. Henry Maudslay’s slide rest lathe for cutting screws from around 1800 represents one of the earliest and, for ages, prevalent approaches to relaxing the separation of control and power mechanisms. As expressed in a state-of-the-art review from 1855, published on the occasion of the World Exhibition in London in 1851, by the end of the 18th century ‘nearly ever part of a machine had to be made and finished to its required form by mere manual labour; that is, we were entirely dependent on the dexterity of the hand and the correctness of the eye of the workman, for accuracy and precision in the execution of the parts of machinery’. However, with the ‘the advances of the mechanical processes of manufacture […] a sudden demand for machinery of unwonted accuracy arose’. Maudslay’s new approach consisted in introducing a ‘slide rest’, i.e., a template whereby the movements of the lathe’s cutting tool would be determined: ‘The principle here alluded to is embodied in a mechanical contrivance which has been substituted for the human hand for holding, applying, and directing the motion of a cutting-tool to the surface of the work to be cut, by which we are enabled to constrain the edge of the tool to move along or across the surface of the object, with such absolute precision, that with almost no expenditure of muscular exertion, a workman is enabled to produce any of the elementary geometrical forms — lines, planes, circles, cylinders, cones, and spheres —with a degree of ease, accuracy, and rapidity, that no amount of experience could have imparted to the hand of the most expert workman.’ (Committee of General Literature and Education, 1855, pp. 238 f.) 89 For an overview of this development, cf. Stuart Bennett’s two volume History of Control Engineering since 1800 (1979, 1993). 320 Cooperative Work and Coordinative Practices This technology was soon applied as ‘part of every lathe, and applied in a modified form in the boring mill, the planing machine, the slotting engine, the drilling machine, &c. &c’ (ibid.) where geometrically accurate lines, planes, circles, cylinders, cones, and spheres were required and gave rise to simple automatic machines for manufacturing parts. ‘Soon after its introduction the slide-rest was made self-acting, that is, its motion along or across the surface to which the tool it held was applied were rendered independent of the attention of the workman in charge of it.’ (ibid.). (Cf. also, Gilbert, 1958). In his discussion of this technology, Hirschhorn makes an important point: ‘From the beginning, automatic machine tools required some form of template so that the eyes and hands of the worker, which once guided the tool, could be replaced by a machine piece. The template became the control element of the machine, determining the feed rate and movement of the tool.’ (Hirschhorn, 1984, p. 22). What Hirschhorn does not say but implies is that a control mechanisms based on a template is of course an pattern transference mechanism in that it transfers its shape or some transformation of its shape. Jacquard’s famous punched-card device for controlling silk-weaving looms (also from the beginning of the 19th century) operated in a similar manner, the string of cards imparting — again in an analog manner — a particular pattern to the fabric (cf. W. English, 1958). An impressive piece of design in its own right, the Jacquard loom is not belittled by observing that it is farfetched and rather fanciful to see in it some kind of anticipation of numerically controlled machinery (e.g., Essinger, 2005). In it own very sophisticated way it basically imparts patterns the way a boot leaves a mark in the snow: by pattern transference. Control devices based on templates have taken many forms and do not really concern us here. It is sufficient to point to the large family of control devices based on a so-called cam, often an irregularly shaped component, that while rotating guide the motions of a ‘cam follower’ in a way similar to the way in which a rotating axis guides the movement of the piston. Compared to other linking mechanisms (gear trains, etc.), the cam provides increased flexibility in developing ‘the form and timing of a particular motion’, since ‘one cam can simply be replaced by another’. ‘In effect’, Hirschhorn states, ‘the control of movement is no longer contained in an array of gears, clutches, and mechanical stops that together make up the structure of the machine, but rather in the cam, which is separate from the body of the machine’ (Hirschhorn, 1984, pp. 22 f.). However, as a control technology camshafts and similar mechanisms again suffered from severe limitations. On one hand, the construction and production of an irregularly shaped object that can impart a sequence of precise movements to a tool posed serious challenges. And on the other hand, and worse, although the cam is physically separate from the machine and can be replaced, it is a part of the energy transmission system in that it both provides the guide for the tool and the force to move it: ‘The linkage between the control and transformation systems of the machine places limits on the economics of cam making. Not only must the cam be powerful enough-that is, of sufficient metallic thickness and mass-to impart the necessary force; it must also be shaped so . as to guide the Chapter 11: Formation and fragmentation 321 tool accurately. Cams tend to be difficult and costly to shape and quick to wear out. Camfollowing machines were thus used mostly for large runs that produced universal pieces such as screws and bolts.’ (Hirschhorn, 1984, p. 23) To put it crudely, mechanical control mechanisms were about as costly to construct and modify as the mechanism they were designed to control. It was for this reason that machine systems remained an economically marginal phenomenon, confined to paper production and similar branches of manufacturing that were producing vast quantities of simple and identical products. The exceptional cases such as transfer lines in automotive industry (car frames, cylinder blocks) were truly impressive but were just that: impressive exceptions in an ocean of highly specialized manual work combined with islands of semiautomatic machine tools and motorized flow lines (Wild, 1972). Machinery only began to spread beyond the confines of mass-production when portable electrical motors began to appear at the end of the 19th century. They were initially applied in machinery such as sewing machines, lathes, drilling machines, and printing presses. But by the 1920s the major part of industry was ‘electrified’. The electrical motor technology offered several advantages but, most importantly, it made it technically and economically feasible to make machines more flexible: ‘as long as power was obtained from a water wheel, a steam engine, or from a large electrical motor that powered a number of devices, the problems of transmitting power to each mechanism and component of a machine required considerable ingenuity. The mechanical problems of transmitting mechanical energy where It was needed to all portions of the machine greatly complicated machine structure.’ (Amber and Amber, 1962, p. 146). With the portable electric motor these almost paralyzing constraints could be relaxed and this made it possible to create machinery that could be modified at lower costs: ‘The mechanic could now place two or more motors in a particular machine. The constraint on machine design was reduced, since different parts could move at different speeds without being connected to the same primary power source. Long gear trains were eliminated. In large machines, independent portable motors could now direct individual segments moving in different planes, eliminating the need for linkages that translated motions in one place to motions in another.’ ‘To change the relative speeds of different machine sections, the mechanic or engineer, instead of stripping the machine and replacing old gears and cams with new ones, need only adjust the relative speeds of the different electric motors. […]. The relaxation of machine constraints opened the way to increasingly general-purpose machines, machines that could be modified at reasonable cost.’ (Hirschhorn, 1984, pp. 19, 21). It was, in practice, only with electromagnetic and electronic control devices (switches, valves, transistors) that the concept of separate control devices began to become articulated, but it was only with the advent of the electronic computer that an economically feasible control technology became available. Machine systems remained an economically marginal phenomenon until — not just the development of the electronic computer in the late 1940s, but until microprocessors around 1980 made it technically and economically feasible not only to design control mechanisms and incorporate them in industrial machinery in the form of 322 Cooperative Work and Coordinative Practices Computer-Numerical Controlled or CNC machinery (as reflected at the time in, e.g., Groover, 1980; Kochhar and Burns, 1983; Koren, 1983) and robotics (e.g., Engleberger, 1980) but also feasible to interlink machines to form automatically coordinated machines systems in the form of Flexible Manufacturing Systems or FMS (e.g., Merchant, 1983) and begin to explore ideas like Computer Integrated Manufacturing or CIM (e.g., Harrington, 1979). In short, it was only with the electronic computer in the form of the microprocessor that an economically viable technical basis for separating control and transformation devices was established. Now, computer technologies did not, of course, originate from the challenges of controlling machines in manufacturing. The stored-program electronic computer was developed for the purpose of automatic processing of massive calculation work. 2.3. The universal control system: The stored-program computer The electronic computer is just as much a machine, i.e., an automatically operating material artifact, as a self-acting mule or a Jacquard loom. It is conceivable that it could be constructed by wheels and gears or electromechanical switches. The difference is that in the former case it is the substantive nature of the moving parts, the interaction of ‘rigid bodies’, that causes the state change; in the electronic mechanism it is the electrical charge that causes the state change. The (enormous) advantage of the latter is first of all that electrons, due to their small mass compared to gears, can travel at a velocity close to the speed of light with minimal expenditure of energy. Where one of Babbage’s gears may have had a mass of, say, 100 grams, or the parts of electromechanical switch circuits in tabulating machines that were in use for most of the 20th century may have had a mass of only 1 gram and hence could work so much faster, electrons have a mass of 9 × 10-28 grams (Goldstine, 1972, p. 144). This of course means that ‘turning a bit’ ‘on’ or ‘off’ requires an insignificant amount of energy. Instead of state changes propagating in a system of interconnected gears and shafts, clouds of electrons are milling about in an equally causal way. In a ‘macroscopic’ mechanism, one cogwheel meshes with another, the movement of the first causing the next cogwheel to change to another position, and a discrete state change has been propagated within the mechanism. Similarly, in an electronic mechanism electrons are amassed at a certain ‘gate’ and, when the charge has reached a pre–specified threshold value, the state of the gate switches. In either case, the pattern of state changes is an observable and highly regular correlation. But just as gears may jam, a wayward cloud of electrons amassing at a particular spot of the circuit (triggered by, say, energetic particles from cosmic radiation) may cause a gate to open and thus turn a bit. Accordingly, a lot of engineering skill goes into the design of both types of mechanism — ‘macroscopic’ as well as ‘microscopic’ — in order to ensure that state changes that are supposed to propagate, and only those, do propagate; in Chapter 11: Formation and fragmentation 323 short, that state changes propagate in a dependable way.90 That is, electronic computers are causal systems just as much as any other machines, from the medieval water clock to the mechanical calculator to the punched-card tabulator. Whether macroscopic or microscopic, the behavior of a machine is a causal process, or a configuration of causal processes, that has been harnessed and thus, under certain conditions, behaves in an extremely regular fashion. What we normally call ‘a computer’ — the Macintosh laptop in front of me, say, or a mainframe computer somewhere in a climate-controlled window-less room — of course consists of a number of machines that are quite tangible: one or more CPUs as well as various specialized integrated circuits such as arithmeticlogic unit, data bus, input and output devices, network connections, etc. In addition, however, ‘the computer’ consists of machines, ‘software programs’, that are just as material as ‘chips’: they are just invisible and intangible, but so are X-rays and TV signals. The computer owes its advantages not just to the enormous speeds afforded by electronics but to the stored program architecture originally outlined by Alan Turing in his famous article on computable numbers (Turing, 1936).91 Programs are treated as data and, when launched, reside in the computer’s storage as — invisible but no less physical — electronic patterns that can be activated and deactivated virtually instantaneously. In short, the computer can be seen as the ultimate control mechanism; or rather, in Turing’s words in a talk on his design for the Automatic Computing Engine (ACE) given in 1947, the computer can ‘imitate’ the control mechanism of any machine: a typesetter, a printing press, a lathe, a machining center, a jukebox. Turing started by referring to his paper ‘On computable numbers’ (1936): ‘Some years ago I was researching on what might now be described as an investigation of the theoretical possibilities and limitations of digital computing machines. I considered a type of machine which had a central mechanism, and an infinite memory which was contained on an infinite tape. This type of machine appeared to be sufficiently general.’ (Turing, 1947, p. 378). He then elaborated on the implications of the stored program architecture: ‘It can be shown that a single special machine of that type can be made to do the work of all. It could in fact be made to work as a model of any other machine. The special machine may be called the universal machine; it works in the following quite simple manner. When we have decided what machine we wish to imitate we punch a description of it on the tape of the universal machine. This description explains what the machine would do in every configuration in which it might find itself. The universal machine has only to keep looking at this description in order to find out what it should do at each stage. Thus the complexity of the machine to be imitated is concentrated in the tape and does not appear in the universal machine proper in any way. 90 As Herman Goldstine puts it, the ENIAC ‘had to operate with a probability of malfunction of about 1 part in 1014 in order for it to run for 12 hours without error. Man had never made an instrument capable of operating with this degree of fidelity or reliability, and this is why the undertaking was so risk a one and the accomplishment so great.’ (Goldstine, 1972, p. 153). 91 ‘In the conventional literature, von Neumann is often said to have invented the stored-program computer, but he repeatedly emphasized that the fundamental conception was Turing’s’ (Copeland and Proudfoot, 2005, p. 114 et passim). 324 Cooperative Work and Coordinative Practices If we take the properties of the universal machine in combination with the fact that machine processes and rule of thumb processes are synonymous we may say that the universal machine is one which, when supplied with the appropriate instructions, can be made to do any rule of thumb process. This feature is paralleled in digital computing machines such as the ACE. They are in fact practical versions of the universal machine. There is a certain central pool of electronic equipment, and a large memory. When any particular problem has to be handled the appropriate instructions for the computing process involved are stored in the memory of the ACE and it is then “set up” for carrying out that process.’ (Turing, 1947, p. 383). The Universal Turing Machine is a mathematical construct, not a real-world machine; it has infinite storage, which no real machine can have, of course. So, while mathematics deals with ‘theorems, infinite processes, and static relationships’, ‘computer science emphasizes algorithms, finitary constructions, and dynamic relationships’. This means that ‘the frequently quoted mathematical aphorism, “the system is finite, therefore trivial,” dismisses much of computer science’ (Arden, 1980, p. 9). That is, the Universal Turing Machine cannot be taken simply as the theoretical model of the computer. However, the concept of the storedprogram computer as ‘universal’ or, better, inexhaustibly malleable, is the key concept in computing technology. The stored-program computer can be reconfigured endlessly, manually or automatically, in response to internal and external state changes. In fact, what we call a computer forms a hierarchy of distinct but interacting control mechanisms: an ‘operating system’ that itself is a hierarchical system of control mechanisms devoted to the managing of data and programs stored in RAM, input and output and external storage devices (‘drivers’), and so on, as well as so–called application programs that may also be hierarchical systems and form hierarchical relations with other application programs. This means that the configuration and reconfiguration of the system of machines constituting the computer, the programs in RAM as well as programs on local harddisks and remote ‘servers’, can be controlled automatically, one machine triggering the execution of another in response to certain conditions, and that human intervention, if required, can be performed semi–automatically or ‘interactively’ as, for instance, a mouse click making one software machine activate another. What is more, also the design and construction of software machines (i.e., programming, compiling, testing) can be performed semi-automatically. This again means that designing and constructing software machines is immensely cheaper and faster than designing and constructing, say, a camshaft or a gear train, or for that matter rewiring the ‘switchboard’ control panel of an punched-card tabulator. And it goes without saying that what applies to designing and constructing software machines applies equally to redesigning and reconstructing software machines. Moreover, of course, software machine design specifications can be copied and transported automatically and at insignificant cost. And, finally, the stored-program technology also makes the construction and modification of largescale machine systems incomparable inexpensive. In short, with the electronic stored-program computer, we have the technology for ‘the production of ma- Chapter 11: Formation and fragmentation 325 chines by machines’ vaguely but perspicaciously anticipated by Marx (1867a, p. 314). However, computing technologies did not come out a box, ready to ‘plug and play’. They do not, first of all, originate from a particular body of mathematical theory; to be sure, their development have depended critically upon a host of mathematical theories (recursive function theory, Boolean algebra, Shannon’s information theory, etc.), but they were not the result of the application of any particular theory. As pointed out by the eminent historian of computation Michael Mahoney, computer science has taken ‘the form more of a family of loosely related research agendas than of a coherent general theory validated by empirical results. So far, no one mathematical model had proved adequate to the diversity of computing, and the different models were not related in any effective way. What mathematics one used depended on what questions one was asking, and for some questions no mathematics could account in theory for what computing was accomplishing in practice.’ (Mahoney, 1992, p. 361). It is no surprise, then, that ‘the computer’ was not ‘invented’: ‘whereas other technologies may be said to have a nature of their own and thus to exercise some agency in their design, the computer has no such nature. Or, rather, its nature is protean’ (Mahoney, 2005, p. 122). It would be more accurate to conceive of this in terms of costs and thus say that computing technology is protean in that the costs of construction and modification of software machines are drastically reduced compared to those of previous machine technologies. Anyway, according to Mahoney, there therefore was a time, a rather long time, ‘when the question “What is a computer, or what should it be”, had no clear-cut answer’ and the computer and computing thus only acquired ‘their modern shape’ in the course of an open-ended process that has lasted decades (Mahoney, 1992, p. 349). And there is no reason why one should assume that the concept of computing as we know it has solidified and stabilized: the jury is still out, as the immense malleability of ‘the computer’ is being explored in all thinkable directions. In other words, it is confused to conceive of ‘the computer’ as one technology. Not only does ‘the computer’ in front of me incorporate a host of technologies, from metallurgy to semiconductor technology to programming languages and operating systems; it can assume an endless range of very different incarnations. It is a protean technology, and how it develops is determined by its applications in a far more radical sense than any other technology. To understand the place of CSCW in this open-ended array of known and possible forms of technology requires that we have an idea of the received concepts of computing: the practical applications for which various computing technologies were developed and that, accordingly, have formed our conceptions of computing and computational artifacts. To do so, I will highlight some of the conceptually distinct forms. 326 Cooperative Work and Coordinative Practices 2.4. Origins of computing technologies in cooperative work Electronic computing technologies initially arose from the development of technologies of large-scale calculation work in science and engineering, on one hand, and in administrative work organizations on the other. The progressive forms of work organization discussed above also appear as a recurring theme in the development of the practices of computing. For a couple of centuries the development of these technologies followed the pattern of development we have dubbed progressive forms of work organization. While this might be taken as proof that these forms are indeed universally necessary forms of development, the banal truth may be simply that managers of large-scale computing work knew of the ‘putting out’ system and the systematic division of labor of the manufactures and applied these as tested technologies. Whatever it is, the major milestones are the following. 2.4.1. Division of ‘mental labor’ In the first year of the French Revolution the new revolutionary regime decided to scrap the received systems of measurement and introduce a new and conceptually coherent system based on the decimal notation. The motive was no secret: it was an intervention to ensure that the myriad of local measurement systems did not pose obstacles for the new regime in its need for raising taxes: ‘At that time, each region was free to establish its own set of measures. Local officials easily manipulated these measures to their own advantage in a number of ways. Commonly, they could keep a large measure to collect taxes of grain and produce but reserve smaller measures for the payment of their own debts’ (Grier, 2005, pp. 33 f.). On top of these mundane motives, however, the new metric system, as it is now known, should be devised and presented in such a way as to demonstrate the grandeur of the revolutionary regime. Hence the new metric system would only allow decimal fractions for the sub-divisions of the units of measure ‘of whatever type’, and accordingly the quadrant of a circle and angular measures were also to be made to conform to this rule. The task of producing these table was assigned to the director of the Bureau du cadastre at the Ecole des ponts et chaussée, Gaspard de Prony, who later related that this implied that ‘all existing trigonometric tables, whether presented in natural or logarithmic form […] became useless; and it was found necessary to calculate new ones’ (de Prony, 1824). Moreover, also in the name of grandeur, de Prony ‘was engaged expressly not only to compile tables which left nothing to be desired about their accuracy, but also to make of them “a monument to calculation the greatest and the most impressive that had ever been executed or even conceived”’. Adding that these were ‘the exact expressions that were used in the brief’ he was given, de Prony specified that this meant that the tables would have to be extended and calculated to 14 or 15 decimal places instead of 8 or 10. Chapter 11: Formation and fragmentation 327 De Prony accepted the assignment ‘unconditionally’ but, realizing that he ‘could not hope to live long enough to finish the project’, he found himself an ‘embarrassment more arduous than [he] could hide’. However, a happy circumstance ‘unexpectedly’ helped him out of this ‘embarrassment’: ‘Having one day noticed, in the shop of a seller of old books a copy of the first English edition 1776, of Smith’s “Treatise on the Wealth of Nations”, I decided to acquire it, and on opening the book at random, I came across the chapter where the author had written about the division of labour; citing, as an example of the great advantages of this method, the manufacture of pins. I conceived all of a sudden the idea of applying the same method to the immense job with which I had been burdened, to manufacture my logarithms as one manufactures pins. I have reasons to believe that, without realising it, I had already been prepared for this realisation from certain parts of mathematical analysis, on which I had then been giving tuition at the École Polytechnique.’ (de Prony, 1824). De Prony may have been exaggerating somewhat in this account, given about 30 years after the event, perhaps as a rhetorical gesture, for the pin manufacture paradigm was very well known in the circles of engineers and scientists he belonged to. In fact, his teacher at École des ponts et chaussée, benefactor, and later immediate superior and predecessor as director at the École, was the very same JeanRodolphe Perronet who had provided the most thorough analysis of the very same pin manufacture in the Normandy that Smith had described (Bradley, 1998). Be that as it may, de Prony certainly applied the principles of manufactures to accomplish the task at hand. In a way, the division of labor had already been applied some fifty years before by Alexis-Claude Clairaut in an effort to calculate the orbit of the comet Halley had identified in 1682 and thereby its next perihelion. The task was massive because the calculation of the orbit of the comet required a solution to a ‘three body problem’, as the orbit is influenced by the gravitational fields of large bodies such as the Sun, Saturn, and Jupiter. Halley did not manage to arrive at a satisfactory solution but Clairaut invented a method of dividing the calculation in such a way that it could be performed in a system of division of labor. Together with two friends, Joseph-Jérôme Lalande and Nicole-Reine Lepaute, Clairaut launched on the massive task. Almost five months later, in November 1758, they were able to publish the prediction that the next perihelion would occur on 13 April 1759 (Grier, 2005). Although the prediction was one month off the mark, the principle of performing massive computations in parallel, as a cooperative effort based on systematic division of labor, was picked up and refined for similar purposes. Lalande himself, from 1759 tasked with calculating astronomical tables published annually by the Académie des sciences under the title Connaissance des temps, employed a small number of skilled ‘computers working out of their own homes’ to do so (Croarken, 2003). In the UK, Nevil Maskelyne, the Astronomer Royal at Greenwich, similarly tasked with the provision of British sailors with a practical technique of determining the longitude, advocated a method (the ‘lunar distance method’) that also required annual calculation and publication of tables (in an Nautical Almanac), and he thus had to devise a computing system that would enable him to accomplish this as a practical task. He was familiar with the form of 328 Cooperative Work and Coordinative Practices work organization that had been developed by Lalande (whom he knew well) and adopted this ‘distributed structure, albeit on a slightly larger scale’: ‘Maskelyne proposed employing a number of computers, each of whom was to undertake the complete set of calculations for specified months of the Nautical Almanac’, and for that purpose he ‘designed a distributed system using relatively skilled workers and which had more in common with cottage industries such as lace making than in the factory manufacture of pins’ (Croarken, 2003, p. 52). Maskelyne provided not only paper and ink, but also ‘computing plans’, that is, instructions that were written on one side of a sheet of folded stationery and that would summarize each step of the calculation. ‘On the other side of the paper he drew a blank table, ready for the computer to complete’ (Grier, 2001, p. 30). De Prony, who knew Maskelyne well and would have been familiar with the British arrangement of mass-calculation work (Grier, 2001, p. 35), did not have to start from scratch. However, whereas the previous experiments in cooperative calculation based on division of labor remained entrenched in artisanal work, as in the example of calculating the perihelion of Halley’s comet, or in the ‘putting out’ system, as in the examples of calculating astronomical tables, de Prony went all the way, so to speak, and adopted the form of full-fledged manufactures. He devised ‘his new manufacture’, as he later called it, as a ‘system of the division of labour’ in three sections: The first was composed of four or five ‘geometricians of very high merit’ that were given the task of choosing the mathematical formulae to be used for calculation and checking; the second section was composed of ‘calculators, who possessed a knowledge of analysis’, and whose task it was to construct the ‘spreadsheets’ to be filled in by the members of the third section. The sheets were divided into 100 intervals, with the numbers of the top line was provided by the ‘calculators’ of the second section. ‘This third group comprised no less than seventy or eighty individuals; but it was the easiest to form, because, as I had foreseen, they did not need, in order to be admitted [to this group] any preliminary instruction; the one essential condition, for their admission [to the group], was for them to know the first two rules of arithmetic […]. The ninety-nine remaining lines were then filled in by means of purely mechanical operations carried out by the 3rd section, each of whom was performing 900 to 1000 additions or subtractions per day, nearly all of whom not having the least theoretical notion on the work which they were doing.’ (de Prony, 1824) The claim that the computers were only required to master addition and subtraction is emphasized by the fact that the ranks of the third group were staffed by former ‘hairdressers’, that is, wig dressers. They were in deep trouble after the aristocratic fashion of wearing wigs suddenly had become a hazardous one (Grattan-Guinness, 1990). De Prony said almost as much when he in his account decades later said ‘that many among them came to seek and find, in this special workshop, a safeguard, a refuge which, happily, was not violated, and that the political circumstances of that time rendered these fully necessary’ (de Prony, 1824). The ‘calculators’ certainly had no prior understanding of interpolation but that did not prevent the cooperative effort from working very efficiently, producing 700 results per day (Grattan-Guinness, 1990). When the project was completed, in Chapter 11: Formation and fragmentation 329 1801, the concerted work of the former hairdressers had produced ‘about 2,300,000 numbers, of which 4 or 500,000 consisted of 14 to 25 digits’ of which ‘99 per cent’ had been calculated ‘by means of a manufacturing procedure’. The tables were ultimately collected in ‘seventeen grand in-folio volumes’ and were deposited at the Paris Observatory. The quality of the work of the calculators was as flawless as it gets. Having explained that ‘All the calculations were done twice: expedite means of verification, but very rigorous, were prepared in advance’, de Prony emphasized that he ‘noticed that the sheets the most exempt from error were, in particular, furnished by those who had the most limited intelligence [education], an existence, so to speak, “automatique”’(de Prony, 1824). The former hairdressers were thus working in an arrangement very similar to that of the workers in the pin manufactures: they mastered their local task fragment but not the larger scheme as conceived by de Prony and his master planners. The historical role of de Prony’s calculation manufacture is similar to Adam Smith’s pin manufacture case, although it has not been as dramatic and hyped. First of all, the cooperative arrangement of calculation organized by de Prony became widely known, not least due to the prominence given to it by Charles Babbage. In 1821 Babbage and his friend John Herschel undertook to produce a set of mathematical tables for the British Nautical Almanac. Organizing the work in accordance with the principles devised by Maskelyne, Babbage and Herschel employed two skilled calculators under the ‘putting out’ scheme. — Now, to understand the work the two men were engaged in, one should be aware that mathematical tables were of enormous practical importance in the emerging industrial economy but were also ridden with calculation and typesetting errors. The Nautical Almanac, a set of tables that from the point of view of maritime safety was ‘crucially significant’, ‘contained at least one thousand errors, while the multiplication tables its computers used sometimes erred at least 40 times per page’ (Schaffer, 1996, p. 277). Herschel later commented that ‘an undetected error in a logarithmic table is like a sunken rock at sea yet undiscovered, upon which it is impossible to say what wrecks may have taken place’ (Swade, 2003, p. 157). In short, assured reliability of calculation was of critical economic and social importance. — So, while Babbage and Herschel in 1821 were engaged in the tedious task of proofreading the manuscripts they had received from their two calculators, it was suggested by one of them, ‘in a manner which certainly at the time was not altogether serious, that it would be extremely convenient if a steam-engine could be contrived to execute calculations for us, to which was replied that such a thing was quite possible, a sentiment in which we both entire concurred’ (Babbage, 1822, quoted in Campbell-Kelly, 1994b, p. 14). The shared sentiment should be understood on this background. At that time, in 1821, Babbage and Herschel knew of de Prony’s work. A few years earlier, in 1817, both of them had signed a letter recommending de Prony’s appointment to the British Royal Society (Bradley, 1998, p. 209), and in 1819, the two of them had visited Paris, and during this trip Babbage was able to inspect, to some extent, the Tables du cadastre that— in spite of having been produced twenty years earli- 330 Cooperative Work and Coordinative Practices er — still had not been printed (for financial reasons). During a visit at the designated publisher of the tables, Didot, Babbage was able to see typeset pages and was given a copy of the section of the sine tables (Schaffer, 1996, p. 278). Anyway, shortly after realizing that calculations could be realized ‘by steam’, Baggage began to design an experimental prototype to demonstrate the feasibility of mechanical production of tables on the basis of the method of finite differences (with which he was ‘completely familiar’ well before his visit to Paris, cf. Lindgren, 1987, pp. 44 f.). We do not know to which extent he in this effort took de Prony’s table manufacture as a guide or model, but when he in June 1822 presented the prototype of the Difference Engine to the public — he did that in an open letter to the resident of the Royal Society to obtain official support (Babbage, 1822) — Babbage used de Prony’s Tables du cadastre as a proof of concept: if a carefully arrangement of workers, based on division of labor, can produce sophisticated mathematical tables of high quality by means of straightforward but repeated addition, then a properly designed machine could do the same, thus making ‘the intolerable labour and fatiguing monotony of a continued repetition of similar arithmetical calculations’ redundant while at the same time reducing the required labor force by about 88 percent. Babbage is often described as one of the earliest pioneers of mechanical computing. His work is of course well-known already and this is not the place for an account and discussion of his impressive oeuvre.92 But from a practical point of view, Babbage’s work on the Difference Engines and the Analytical Engine could be seen as a wasted effort. In spite of the large sums that were invested in their design and construction, none of the projects were brought to completion. However, his work is of course of great historical interest in its own right. First of all, it had direct influence on the design of a series of difference engines that were built and used over the next years by engineers such as Scheutz, Wiberg, Grant, and Hamann, and in contrast to Babbage’s designs, some of these machines were actually used (Lindgren, 1987). But on balance, the approach developed by Babbage turned out to be of marginal practical utility (M. R. Williams, 2003). Or in the words of Alan Bromley, ‘That these were not extensively used or developed, despite the apparent complete success of the Wiberg machine, indicates that the entire idea was not well judged. The sub-tabulation task, though laborious, was not the dominant mathematical task in the preparation of tables nor, with adequate organization and management, was it of overwhelming practical importance’ (Bromley, 1990, p. 96). One is thus led to the conclusion that the ‘fruits of Babbage’s considerable genius’ were ‘effectively wasted as far as practical influence is concerned’ (ibid., p. 97). One should of course not underestimate the moral example of his projects. They demonstrated to computing researchers in the 20th century (such as Howard 92 For general descriptions of the development of Babbage’s work, cf. his autobiography (Babbage, 1864, chapters V, VII, and VIII), the host of studies of history of technology devoted to Babbage (e.g. Collier, 1970; Lindgren, 1987; Bromley, 1990; Schaffer, 1996), as well as a few reliable popular biographies (e.g., Swade, 2000). Chapter 11: Formation and fragmentation 331 Aiken and Vannevar Bush) that complex computation by means of automatic digital artifacts was feasible. However, the inspiration has never been transformed into anything technologically specific. In fact, it is when ‘we come to examine the facilities available for programming the Analytical Engine that Babbage’s designs begin to look strange to modern eyes’ (Bromley, 1990, p. 87). ‘The conclusion seems inescapable that Babbage did not have a firm command of the issues raised by the user-level programming of the Analytical Engine. It would be quite wrong to infer that Babbage did not understand programming per se.’ In so far as programming is concerned, his focus was what we now call microprogramming and it was from this base that Babbage explored the ideas of user-level programming. However, ‘The issues of data structuring simply did not arise at the microprogramming level. There is some evidence to suggest that Babbage’s ideas were moving in the directions now familiar in connection with the control mechanisms for loop counting in user-level programs. Had an Analytical Engine ever been brought to working order, there can be no doubt that Babbage’s programming ideas would have been developed greatly.’ (Bromley, 1990, p. 89) In other words, if one considers the contribution of the Babbage engines narrowly from the point of view of the sophisticated technicalities and divorced from its use, one is likely to miss the fact that the technology he developed was not used and that he did not arrive at the point where use issues did arise. The real importance of de Prony’s example is not its probable impact on de development of Babbage’s engine designs. It was, in the words of Babbage, ‘one of the most stupendous moments of arithmetical calculation which the world has yet produced’ (Babbage, 1822, p. 302) and in his On the Economy of Machinery and Manufactures he not only gave a detailed account of de Prony’s accomplishment but he did so under the heading ‘On the division of mental labour’ (Babbage, 1832, §§ 241-247), stating the conclusion to be drawn from de Prony’s example sharply in the opening sentence of the chapter: ‘We have already mentioned what may, perhaps, appear paradoxical to some of our readers that the division of labour can be applied with equal success to mental as to mechanical operations, and that it ensures in both the same economy of time.’ (Babbage, 1832, § 241). To see the historical importance of this, one should know that it was this book, not the illfated design projects, that defined Babbage’s reputation among his contemporaries, and it was reprinted many times and translated to a large number of European languages. He was, in fact, better known as an economist than as a technologist. His book was based on extensive and conscientious field work in the textile manufacturing districts of England and on the Continent undertaken in course of the 1820s, and as pointed out by the historian Campbell-Kelly in his introduction to Babbage’s autobiography (1994b), the Economy of Machinery and Manufactures is generally seen as a ‘being in the direct line of descent’ from Adam Smith’s Wealth of Nations to Frederick Winslow Taylor’s Principles of Scientific Management (1911). Accordingly, to the generations of scientists, economists, managers, and workers who read Babbage’s On the Economy of Machinery and Manufactures, de 332 Cooperative Work and Coordinative Practices Prony’s example was the ‘proof of concept’ that the principle of division of labor, which they of course knew from Adam Smith and from their daily work, could be applied equally well to ‘mental labour’. Cooperative work based on advanced division of ‘mental labor’, as devised by de Prony in the wake of the French Revolution, became wide-spread in course of the 19th and early 20th centuries. It not only became the standard way of handling the increased load of scientific and engineering calculation (Grier, 2001), but it also, with the rise of large-scale financial and industrial corporations, became the predominant way of organizing administrative work in such settings. That is, the real importance of de Prony’s example lies in the model it provided for the organization of mental work in the next two centuries: the organization of ‘human computers’ and ‘calculators’ in scientific and engineering laboratories, in insurance companies and accounting offices, in inventory management and production planning in manufacturing, and so forth. (For a description of a classical case, the Railway Clearing House, cf. CampbellKelly, 1994a). Babbage’s direct assault on the mechanization of mental labor came to naught. It was de Prony’s scheme that ruled the day for more than 150 years. For a century after de Prony’s calculation manufacture and for more than half a century after the difference engines designed on the Babbage model, calculation work remained strictly ‘manual’, as one is awkwardly tempted to call it, meaning that all operations are performed by mind and hand, assisted by the use of pen and paper and perhaps an abacus or a slide rule. The technology of mechanical calculation machines only matured in the course of the 19th century and only matured at a glacial speed. For although mechanical calculating machines date back to the 18th century (e.g., Schickard, ca. 1620; Pascal, 1642; Leibniz, 1674), we should not (again) be misled by the chronology of inventions and disregard technology in actual use. As pointed out by a historian of computing, ‘Mechanical calculating machines were essentially useless toys during the first two centuries of their development. The level of [metal-working] technology of the day guaranteed that any attempt to produce a reliable, easy to use instrument was doomed to failure.’ (M. R. Williams, 1990, p. 50). The ‘first machine that can be said to have been a commercial success’ was a calculator created by Thomas de Colmar in 1820, but the technology of mechanical calculation only stabilized late in the 19th century with the development of the Baldwin-Odhner calculator (patented 1875) that offered a practical and robust solution to the carry issue (a variable-toothed gear). The most famous design based on this technology was perhaps the calculators produced by the Brunsviga company in Germany from 1892. The scope and level of automatic control remained quite narrow and low, however; restricted to, for instance, control of carry operations in addition tasks. As in the cotton trades one hundred years earlier, mechanization of ‘mental labor’ was an incremental process. Mechanical calculation machines were operated in isolation from each other, by the individual part-worker, as a means of speeding up the tedious task of doing massive arithmetical tasks. The reason for this is that the level of automatic con- Chapter 11: Formation and fragmentation 333 trol of operations was so rudimentary (automatic carry was the overwhelming issue) that one can safely say that ‘In in these machines the control function was provided by the human operator’ (Bromley, 1990, p. 59). They were little more than sophisticated tools. In these conditions, mechanical calculation machines did not permit the operator to escape the subjection to specialization in performing a specific operation (or a narrow range of related operations) and take on the role of a supervising the operations of the machine. The devices were simply incorporated into the received division of labor as a means for increasing the speed at which individual operations could be performed. 2.4.2. Mechanization of ‘mental labor’ Calculation in administrative work began to become mechanized around the beginning of the 20th century, with the invention and dissemination of punched-card tabulators (and associated punching equipment, sorters, printers, etc.). Tabulating machines were soon equipped with plugboards or switchboards, so that the machines could be reconfigured to handle other tasks and card formats. Invented for use in the processing of massive statistical data (the US census 1890), punched-card machinery quickly became appropriated and used by railroad and utilities companies as well as by manufacturers and government agencies. However, because of their costs, punched-card tabulators were generally confined to use in settings such as these that were in need of (or could exploit) large-scale data processing: ‘Punched-card machinery was expensive to rent and consequently was only used, at first, by very large organizations that could make good use of its ability to make short work of a large volume of transactions; the needs of small businesses could be met adequately by less automatic but lowercost bookkeeping machines, such as those made by Burroughs. The Hollerith [punched-card] machines, however, arrived at a critical period in the development of large-scale American enterprise; it was during this period in the late nineteenth and early twentieth centuries that much of modern business accounting practice came into existence, particularly cost accounting in manufacturing.’ (Campbell-Kelly, 1990, p. 145). As large-scale economic organizations evolved, the use of punched-card tabulating machinery became widespread. By 1913 a journalist reported that ‘the system is used in factories of all sorts, in steel mills, by insurance companies, by electric light and traction and telephone companies, by wholesale merchandise establishments and department stores, by textile mills, automobile companies, numerous railroads, municipalities and state governments. It is used for compiling labor costs, efficiency records, distribution of sales, internal requisitions for supplies and materials, production statistics, day and piece work. It is used for analyzing risks in life, fire and casualty insurance, for plant expenditures and sales of service, by public service corporations, for distributing sales and cost figures as to salesmen, department, customer location, commodity, method of sale, and in numerous other ways. The cards besides furnishing the basis for regular current reports, provide also for all special reports and make it possible to obtain them in a mere fraction of the time otherwise required.’ (quoted in Campbell-Kelly, 1990, p. 145). That is, with the use of tabulating machinery important sections of administrative work assumed the character of the machine-based factory. 334 Cooperative Work and Coordinative Practices It is of some relevance here to note that, in some instances, the technology was used also in the coordination of cooperative work, in a manner that in some ways anticipates kanban systems, as opposed to off-line administration of work settings and processes. In the 1920s and 1930s ‘the increasing variety of styles, colors, and options began to slow production and delay delivery’, and ‘automobile companies turned to the use of tabulator machinery to overcome these delays’ (Norberg, 1990, p. 774). Norberg describes an innovative use of tabulating machinery in Chrysler Corp. for purposes of coordination: ‘Upon receipt of a dealer’s order, two cards were punched with the essential information supplied by the dealer: routing, region, district, dealer’s name, order number, item number, model, body type, paint, trim, wheels, transmission, radio, heater, and other options. One card went to the equivalent of a production planning office where a “Daily Master Building Schedule” was prepared, and one went to the car distribution department where it was filed according to region and dealer. Multiple copies of the production card went to the various inventory-control points for parts, while several copies stayed with the car as it was constructed. When the car reached the shipping department, one of the last cards remaining was checked to see that the order was correct. If so, the car was shipped and the dealer was notified when to expect it.’ (Norberg, 1990, p. 774) In this case the punched-card is not simply as record of an event to be processed at a later stage for secondary use, e.g., for statistical purposes, for purposes of payment, etc. The card is a mechanically generated coordinative artifact that provides the various stations in the large-scale network of activities with appropriate information about the particular order. Like desk-top calculators before them, tabulating machinery was and remained stand-alone machines, with operators supervising the operation of the individual machines and handling the transfer of cards between machines: punchers, sorters, tabulators, printers. On the basis of stacks of discrete cards a higher degree of automatic integration of operations was not feasible, but the technology remained in use until the cost of electronic computing made it a viable option for ordinary work settings to move beyond the confines of punched-card tabulator technology. By the late 1960s traditional punched-card machines had effectively gone out of production, and by the late 1980s punched cards had all but vanished (CampbellKelly, 1990, p. 151). So, although component technologies of punched-card tabulating were appropriated and used in the first generations of electronic computers (punched cards, card readers, printers, plugboards, etc.), the electronic digital computer did not grow out of this technology, nor did it grow out of the needs of administrative work. The first generations of electronic digital computers were designed and built for massive scientific and engineering calculation. While administrative work became mechanized, the mechanization of scientific and engineering calculation, remained sporadic and fragmentary. This lasted until the Second Work War, still relying on desktop calculators. In a few large-scale research settings that could afford the cost, punched-card tabulators were appropriated for the purposes of scientific and engineering calculation (e.g., Snedecor, 1928; W. J. Eckert, 1940; McPherson, 1942), but in general human computers Chapter 11: Formation and fragmentation 335 were still tasked with calculating by rote in accordance with a systematic method supplied by an overseer. Donald Davies, who was to play a key role in the development packet-switched digital networks, recalls from his work as a young scientist in the UK during the Second World War: ‘The Second World War saw scientific research projects of a size and complexity that reached new levels. Underlying much of the work were complex mathematical models, and the only way to get working solutions was to use numerical mathematics on a large scale. In the Tube Alloys project, for example, which became the UK part of the Manhattan Project to make a fission bomb, we had to determine the critical size of a shape of enriched uranium and then estimate mathematically what would happen when it exploded. For this problem we used about a dozen “computers” — young men and women equipped with hand calculators (such as the Brunsviga). These human computers were “programmed” by physicists like myself.’ (Davies, 2005, p. vii). This setting was not a unique. As Davies puts it, the ‘same story, with different physics and different mathematics, was repeated in many centres across the United Kingdom’. In the words of Jack Copeland, ‘The term “computing machine” was used increasingly from the 1920s to refer to small calculating machines which mechanized elements of the human computer’s work. For a complex calculation, several dozen human computers might be required, each equipped with a desktop computing machine.’ (Copeland, 2006b, p. 102). However, just as the vastly increased scope of administrative data processing had put the calculation manufacture form of work organization under increasing pressure and thus engendered the rapid growth of punched-card technologies, the scale of calculations required in modern science and engineering caused similar tensions to arise: ‘By the 1940s, […] the scale of some of the calculations required by physicists and engineers had become so great that the work could not easily be done with desktop computing machine. The need to develop high-speed large-scale computing machinery was pressing.’ (Copeland, 2006b, p. 102). In short, in the domains of science and engineering too, cooperative calculation work, organized on the de Prony model, had exceeded its capacity for further development. In sum, it was the problems facing the cooperative efforts of human computers in science and engineering that motivated the first significant steps towards electronic digital computers, in particular the Colossus, designed by Thomas Flowers and Max Newman at Bletchley Park in the UK during 1943 for breaking the encrypted messages produced by the German Geheimschreiber, a family of sophisticated teletype cipher machines used for strategic communication within the Nazi military leadership (Copeland, 2006a; Gannon, 2006),93 and the ENIAC, designed in the US in 1944-45 for calculating projectile trajectories (Goldstine, 1972; Van der Spiegel, et al., 2000). The need for similar applications motivated the subsequent series of experimental stored-program computers such as, in the US, 93 The British government kept the very existence of the Colossus secret until 1975, and its function was not publicly known until 1996 when the US Government declassified documents, written by US liaison officers at Bletchley Park during the war, in which the function of the Colossus was described. However, a ‘vital report’ (Good, et al., 1945) was only declassified in June 2000 (Copeland, 2006c). 336 Cooperative Work and Coordinative Practices • the EDVAC, 1945-52, designed by John Mauchly, Presper Eckert, and John von Neumann (von Neumann, 1945; J. P. Eckert, 1946); • the Princeton IAS computer, 1946-52, designed by von Neumann (Aspray, 2000), etc., and in Britain, • the ACE, 1945-50, designed by Turing and Wilkinson (Turing, 1945; Turing and Wilkinson, 1946-47; Turing, 1947; Copeland, 2005); • the Manchester Mark I, 1946-48, designed by Newman (Napper, 2000); and • the EDSAC, 1946-49, designed by Maurice Wilkes. (For general accounts of these efforts, cf. Augarten, 1984, Chapters 4-5; Campbell-Kelly and Aspray, 1996, Chapter 4) If we take Turing’s ACE as an example, the motivation was clearly laid out. In his proposal, written towards the end of 1945. Turing opened the report by stating: ‘Calculating machinery in the past has been designed to carry out accurately and moderately quickly small parts of calculations which frequently recur. The four processes addition, subtraction, multiplication and division, together perhaps with sorting and interpolation, cover all that could be done until quite recently […]. It is intended that the electronic calculator now proposed should be different in that it will tackle whole problems. Instead of repeatedly using labour for taking material out of the machine and putting it back at the appropriate moment all this will have to be looked after by the machine itself. This arrangement has very many advantages. (1) The speed of the machine is no longer limited by the speed of the human operator. (2) The human element of fallibility is eliminated, although it may to an extent be replaced by mechanical fallibility. (3) Very much more complicated processes can be carried out than could easily be dealt with by human labour. Once the human brake is removed the increase in speed is enormous.’ (Turing, 1945, p. 371). The same motivation was underscored when Charles G. Darwin, the director of the UK National Physical Laboratory, in April 1946 wrote a memorandum in which he argued the case for building the computer proposed by Turing: ‘In the past the processes of computation ran in three stages, the mathematician, the [human] computer, the machine. The mathematician set the problem and laid down detailed instructions which might be so exact that the computer could do his work completely without any understanding of the real nature of the problem; the computer would then use the arithmetical machine to perform his operations of addition, multiplication, etc. In recent times, especially with use of punched card machines, it has been possible gradually for the machine to encroach on the [human] computer’s field, but all these processes have been essentially controlled by the rate at which a man can work.’ (Darwin, 1946, p. 54). Darwin went on by stressing that ‘The possibility of the new machine started from a paper by Dr. A. M. Turing some years ago when he showed what a wide range of mathematical problems could be solved, in idea at any rate, by laying down the rules and leaving a machine to do the rest’ (Darwin, 1946, p. 54). Computing technology as represented by these pioneering calculating machines were designed to eliminate the ‘human brake’, that is, the cooperative work of human computers and punched-card operators, just as automatic machine systems a century previously, in other domains of work, had eliminated the coopera- Chapter 11: Formation and fragmentation 337 tive work of workers in paper mills, etc., while of course constituting cooperative work of an entirely different sort, performed by machine operators and technicians. The first electronic digital computers such as the Colossus and the ENIAC were not stored-program computers. They were specifically designed for performing massive calculations, not as general purpose computers. Thus, to facilitate configuration and reconfiguration the machines were equipped with plugboards similar to those used in punched-card tabulators and, in the case of the Colossus, also switches. However, the configuration work was tedious and the cost in terms of time significant. For example, configuring the ENIAC by plugging cables, ‘its users were literally rewiring the machine each time, transforming it into a specialpurpose computer that solved a particular problem’; and, consequently, it took ‘up to two days’ to configure the ENIAC to solve a new problem, which it might then solve in a minute (Ceruzzi, 1990, p. 241). Moreover, the ENIAC had been designed for calculating projectile trajectories and was not particularly suited for other types of massive calculation such as solving partial differential equations (Campbell-Kelly and Aspray, 1996, p. 91). The impetus to develop storedprogram computers came from such limitations. The stored-program digital computer technology (e.g., EDVAC and ACE) had been developed, as a ‘universal’ or ‘general’ technology of large-scale calculation machinery that was far less expensive and far more flexible than building series of specialized calculation machine systems. But it was then — in one of those lateral shifts in which a technology developed for one domain of work is picked up and appropriated for another domain — gradually and hesitatingly transformed and appropriated for administrative purposes. The application of stored-program electronic computers in work settings only began in the 1950s. The first ‘business’ application, a payroll program, ran on 12 February 1954, on the then newly finished LEO computer (based on the Manchester Mark I architecture), calculating the wages of bakery staff of the Lyons teashop chain in the UK (Ferry, 2003). This application is typical for the use of electronic computers for business purposes, that is, for economic, commercial, organizational, managerial, etc., purposes (beyond applications of scientific calculation): batch processing of large numbers of transaction records. The stored-program electronic computer was appropriated for business administration purposes as a substitution technology; that is, the new computers were designed for automating the work of the central computing departments of largescale organizations, replacing entire batteries of punched-card tabulators by a single computer such as the IBM 1401. Most of the computer systems installed in commerce and government during the 1950s and 1960s were ‘simply the electronic equivalents of the punched-card accounting machines they replaced.’ (Campbell-Kelly and Aspray, 1996, p. 157). The punched-card technology had helped to shape the organization of business, and by the 1950s and 1960s ‘the highly centralized accounting systems of industry were very much geared to what was technically achievable with the commercially available machines, and a generation of accountants be- 338 Cooperative Work and Coordinative Practices tween the two world wars grew upon a diet of the standard textbooks on mechanized accounting. When computers became available in the 1950s and 1960s, they tended to be used at first as glorified electric accounting machines and were simply absorbed into old-fashioned accounting systems.’ (Campbell-Kelly, 1990, pp. 146 f.). What characterized this computing technology was automatic processing (recording, sorting, merging, aggregating, calculating, printing) of data concerning economic activities: statistical analysis (actuarial data, sales analysis), invoicing, sales reports, payroll, inventory control, financial reporting. That is, this was a technology developed and used for administrative and logistical purposes in ordinary business settings (payroll calculation, production planning). The computer systems were used for performing various house-holding tasks in organized cooperative work settings but the interdependent activities of the cooperating workers were neither facilitated by the system, nor mediated in and through the system, nor regulated by the system: one worker’s actions were not effectuated and propagated to other workers by means of the system. The system remained ‘outside’ of the practices and settings whose economic transactions it was processing. 2.5. Facilitation of cooperative work: Real-time-computing Parallel to the development of these technologies of automatic handling of administrative calculation, an entirely different computing technology was being developed that directly addressed the facilitation of cooperative work, namely ‘online’ ‘real-time’ computing systems such as air defense systems (SAGE) and airline reservation systems (SABRE). With this technology computational machine systems were constructed that would constitute the common field of work of multiple cooperating actors interacting ‘in real time’. 2.5.1. Project Whirlwind The increased role of airplanes in warfare in World War II caused two bottlenecks: testing new airplane designs and training crews for them were costly and caused intolerable delays: ‘in 1943 it was taking far too much time and money to train flight crews to man the more complex, newer warcraft in production, and it was taking far too much time and money to design high-performance airplanes’ (Redmond and Smith, 1980, p. 1). The obvious path of developing a particular flight simulator for each particular airplane model was not sustainable, as it would lead to increased costs of ‘providing a new and different flight trainer for each warplane model in combat use’. In 1943, this prospect led researchers at MIT’s Servomechanisms Laboratory together with US Navy planners to develop the alternative strategy of developing configurable flight simulators that could match the flight characteristics of any particular airplane design: ‘a protean, versatile, master ground trainer that could be adjusted to simulate the flying behavior of any one of a number of warplanes’ (Redmond and Smith, 1980, p. 2). The project, Chapter 11: Formation and fragmentation 339 named Whirlwind, was undertaken in 1944 under the leadership of Jay W. Forrester.94 To realize the idea of a ‘protean and versatile’ simulator, the simulator should incorporate a computer and, what is more, a computer with the capacity to process incoming data and solve the system of differential equations at a rate that was sufficient to match the rate of external state changes. That is, the computer should be able to handle external events rapidly enough to be ready for the next external event or ‘in real time’. Forrester initially opted for a design based on an analog computer, but it was clear that an analog computer ‘would not be nearly fast enough to operate the trainer in real time’ (Campbell-Kelly and Aspray, 1996, p. 159). However, in the summer of 1945 he learned that digital computer technology was a viable option. He learned this from another MIT student, Perry Crawford, who had developed some initial concepts for real-time process control based on digital electronic calculating systems. Noting that it had ‘recently’ been proposed at MIT ‘that electronic calculating systems can perform a valuable function in fire-control operations’, his thesis set out to ‘describe the elements and operation of a calculating system for performing one of the operations in the control of anti-aircraft gunfire, which is, namely, the prediction of the future position of the target.’ (Crawford, 1942, p. 1). At the time, the control systems for gun control and so on were still invariably based on mechanical or electromechanical technologies. What Crawford suggested was that the mathematical functions could be modelled in a digital computer that could thereby be made to control real-world processes such as tracking a moving target: ‘Crawford was the first person to fully appreciate that such a general-purpose digital computer would be potentially faster and more flexible than a dedicated analog computer’ (Campbell-Kelly and Aspray, 1996, p. 160). Crawford explained all this to Forrester in 1945 and, as Forrester later put it, it ‘turned on a light in [my] head’ (ibid.). The stored-program computer technology was made the foundation of the subsequent development work in Project Whirlwind (by virtue of access to the ongoing EDVAC design work, cf. Goldstine, 1972), and as the project progressed, the objective of building a flight simulator receded into the background; instead, the effort focused increasingly on the challenge of building a real-time digital computer system (Crawford, 1946). ‘The two young men quickly realized that they would not be developing simply a sophisticated flight trainer. Instead, they had stumbled onto a design concept so fundamental that its generality of application was almost staggering to contemplate’ (Redmond and Smith, 1980, p. 217). 94 The following account of Whirlwind and its legacy, the SAGE system and beyond, is based on the insightful studies by O’Neill (1992) and by Campbell-Kelly and Aspray (1996). The two volumes by Redmond and Smith (1980, 2000) offer a detailed and accurate account; however, the information that is relevant from the perspective of technology-in-practice has to be dug out from an account that is overwhelmingly focused on issues of research governance and project management; this makes this major piece of research less immediately useful for researchers with CSCW or HCI interests and concerns. 340 Cooperative Work and Coordinative Practices A major technical issue in developing Whirlwind was the speed and reliability of computer memory. Consequently, significant effort was devoted to developing the new technology of magnetic core memory, based on a web of tiny magnetic ceramic rings. Core memory technology, which did not become operational until the summer of 1953, made Whirlwind ‘by far the fastest computer in the world and also the most reliable’ (Campbell-Kelly and Aspray, 1996, p. 167). It was a ‘monumental achievement’ (O’Neill, 1992, p. 13).95 The operational speeds offered by core memory was further underscored by the development for Whirlwind of ‘the intricate details of “synchronous parallel logic” — that is, the transmitting of electronic pulses, or digits, simultaneously within the computer rather than sequentially, while maintaining logical coherence and control. This feature accelerated enormously […] the speeds with which the computer could process its information.’ (Redmond and Smith, 1980, p. 217). Whirlwind was also ‘first and far ahead in its visual display facilities’ which, among other things, facilitated the ‘plotting of computed results on airspace maps’ (Redmond and Smith, 1980, p. 216). Complementary to this feature, was a ‘light gun’ with which the operator could select objects and write on the display: ‘As a consequence of these two features, direct and simultaneous man-machine interaction became feasible’ (ibid.). That the Whirlwind research continued and eventually became massively funded despite the fact that no stored-program digital computer existed at the time (and would not exist until the ‘Manchester Baby’ ran its first test 21 June 1948) was due to external events in 1949. The US intelligence services revealed that the Soviet Union had exploded a nuclear bomb in August that year and furthermore possessed bomber aircraft capable of delivering such weapons at targets in the US. Quickly a committee was put to work to evaluate the implications for US airdefense. In 1950, the committee concluded that the existing air defense system was wholly inadequate for the current situation. It candidly compared the existing system ‘to an animal that was at once “lame, purblind, and idiot-like”’, adding, in order not to leave readers guessing, that ‘of these comparatives, idiotic is the strongest’ (ADSEC Report, October 1950, quoted in Redmond and Smith, 1980, p. 172). The problem was basically that the coordination effort of the existing system severely limited its capacity. As O’Neill explains the conundrum: ‘In the existing system, known simply as the “manual system”, an operator watched a radar screen and estimated the altitude, speed, and direction of aircraft picked up by scanners. The operator checked the tracked plane against known flight paths and other information. If it could not by identified, the operator guided interceptors into attack position. When the plane moved out of the range of an operator’s radar screen, there were only a few moments in which to “hand-over” or transfer the direction of the air battle to the appropriate operator in another sector.’ (O’Neill, 1992, p. 15). 95 Together with printed circuits, core memory made possible the mass production of computers such as the IBM 1401, announced in October 1959. Chapter 11: Formation and fragmentation 341 This cooperative work organization, the ‘manual system’, did not scale up to meet the challenge of large numbers of aircraft with intercontinental reach and carrying nuclear weapons. The problem was systemic (Wieser, 1985). In the words of O’Neill again: ‘In a mass attack, the manual handling of air defense would present many problems. For example, the manual system would be unable to handle detection, tracking, identification, and interception for more than a few targets in the range of anyone radar. The radar system did no~ provide adequate coverage for low altitude intrusion. The way to get around the lack of low altitude surveillance was the “gap filler” radar. Because this “gap filler” radar was limited to a few tens of miles, the system required more frequent hand-overs, and further taxed the manual-control system and its operators by reducing the time available to intercept an attacking bomber.’ (O’Neill, 1992, pp. 1516). It was concluded, then, that a radical transformation of the organization of the airdefense organization and its technical infrastructure was required. The new air defense system that was eventually built and was named Semi-Automatic Ground Environment, or SAGE, was divided into 23 Direction Centers distributed throughout the USA. Each of these centers would be responsible for monitoring the airspace of the sector and, if required, for directing and coordinating military response. The work of each Direction Center was supported by a high-speed electronic digital processing machine that would receive and process data from radar sites via a system of hundreds of leased telephone circuits. In 1950 it was decided that the SAGE computer system would be based on the Whirlwind design. Whirlwind was, of course, an experimental system and was not fit for production. An engineered version of Whirlwind was developed which was initially simply known as Whirlwind II, but then named XD-1 until it finally, as a production version manufactured by IBM, was renamed AN/FAQ-7 (or the Q-7 as it was often called). ‘The direction center itself was to be semiautomatic; that is, routine tasks would be done automatically under the supervision of operators. A high-speed digital computer would collect target reports from the radar network, transform them into a common coordinate system, perform automatic track-while-scan […], and compute interceptor trajectories. Operators filtered the radar data, had override control (i.e., could initiate or drop tracks), performed friend-or-foe identification function, assigned interceptors to targets, and monitored engagements through voice communication with the interceptor pilots.’ (Wieser, 1985, p. 363) To determine the feasibility of the plan, a ‘computer-controlled collisioncourse interception’ test was undertaken on 20 April 1951 above Bedford, Massachusetts. According to the test report by C. Robert Wieser dated 23 April 1951, ‘three successive trial interceptions were run with live aircraft under control of the WWI [Whirlwind I] computer’ (quoted in Redmond and Smith, 2000, p. 1). The pilot of the intercepting aircraft reported that from a distance of about 60 kilometers (40 miles) he was brought to within 1,000 meters of his target. Three days later it was decided to build a prototype of the SAGE system, ‘an elaborate, multiradar experimental system tied into Whirlwind I’ (Redmond and Smith, 2000, p. 2). The development of this experiment prototype, called the Cape Cod system, proceeded in an entirely iterative manner. Robert Wieser, who was deeply in- 342 Cooperative Work and Coordinative Practices volved in the development of the Cape Cod prototype as an engineer recalls that the ‘development of the new concept and its embodiment in the Cape Cod System relied heavily on iterative cycles of experiment-learn-improve. […] However inelegant, the approach worked very well. The ever-present realism of radar clutter, telephone-line noise, and limited computer memory drove the development pace faster than a mathematical analytical approach could ever have done’ (Wieser, 1985, p. 364). The Cape Cod system was based on the ‘engineered version’ of Whirlwind, the XD-1, and was ready for experiments in 1952. It supported 30 air-force operators working next to each other at consoles equipped with CRT displays on which digitized radar data could be selected for analysis by means of a light pen (CampbellKelly, 2003, p. 37). In March 1953, Robert Wieser gave a talk to visitors to a Cape Cod demonstration. Introducing them to what they were about to see, he said: ‘The radar data is fed into the Whirlwind I computer at the Barta Building in Cambridge, which processes the data to provide 1) vectoring instructions for mid-course guidance of manned interceptors and 2) special displays for people who monitor and direct the operation of the system. [¶] In processing data, the computer automatically performs the track while-scan function, which consists of l) taking in radar data in polar coordinates, 2) converting it to rectangular coordinates referred to a common origin, 3) correlating or associating each piece of data with existing tracks to find out which pieces of data belong to which aircraft, and 4) using the data to bring each track upto-date with a new smoothed velocity and position, and 5) predicting track positions in the future for the next correlation or for dead reckoning if data is missed. Once smoothed tracks have been calculated, the computer then solves the equations of collision-course interception and generates and displays the proper vectoring instructions to guide an interceptor to a target. This process is not, however, wholly automatic. The initiation of new tracks can be done automatically or manually, or both methods can be used, each in different geographical areas of the system. Also the decision as to which aircraft tracks are targets and which tracks are interceptors is made by people and inserted manually into the machine by means of a light gun. The light gun is a photocell device which is placed over the desired blip on the display scope and then sends a pulse into the computer to indicate to the computer that action (for example, “start tracking”) is to be taken on that particular aircraft. The action which the machine takes is defined by manually setting a selector switch, which the computer automatically senses and interprets (for example, “handle this aircraft as an interceptor”). The human beings make decisions and improvise while the computer handles the routine tasks under their supervision. In order to facilitate human supervision, a rapid, flexible display system is required. The principal means of display is the cathode-ray tube, which can accept information very rapidly and present both symbols and geographical positions of aircraft. Flexibility is achieved by programming the computer to display various categories of information on different display cables. The human operator can switch these cables at his scope and thus select at any time the type of information (or combination of types) which he wishes to observe.’ (Wieser, 1953, p. 2). From the beginning, that is, the system was planned as ‘semi-automatic’, that is, the computer system was designed to work in a mode radically different from the automatic calculations for which other contemporary computers were being designed. The Cape Cod prototype and the ultimate SAGE computer systems (the Q-7s) were deliberately designed for interactive computing, complete with real- Chapter 11: Formation and fragmentation 343 time computing, graphical CRT displays, handheld selection devices (light guns, joysticks), direct manipulation. The SAGE system was an enormous machine system, a system of 23 interconnected computers (and an equal number of backup computers) connected to a vast array of radar stations, intercept fighter aircraft, etc., that afforded a large-scale cooperative work effort encompassing a distributed ensemble of 2,300 operators. Again, O’Neill’s account is very informative about the actual work at the centers: ‘As many as 100 Air Force personnel used a SAGE computer in a single facility at the same time. They used an assortment of equipment to communicate to the computer, such as cathode ray tube (CRT) displays, keyboards, switches, and light guns. This equipment provided the operators with the information they needed to make decisions and provided them a way to send commands to the computer. The SAGE system was designed to make use of consoles with CRTs for displaying the tracks of radar-reported aircraft and providing visual maps to improve comprehension. Computer programs controlled the beam that created a track’s image on the CRT surface by supplying the coordinates needed in drawing and refreshing the image. Light guns allowed the operators to aim at an unidentified plane image on a screen and the system would then give information about the specified image. If the operator determined that a plane was hostile, the system would tell the operator which of the many interceptors or guided missiles on hand were in the best position-to intercept it. The operators filtered the radar data, had the ability to override, performed the friend-or-foe identification function, assigned interceptors to targets, and monitored engagements through voice communication with the interceptor pilots. SAGE required visual displays, real-time responsiveness, and communication between computers. These are all elements of interactive computer use. From the beginning, SAGE was designed to replace the manual system only partially. It was semiautomatic; a human element remained an important part of the system. The splitting of tasks was seen as a good approach to the problem; machines were necessary because human operators working without computers could not make the necessary calculations fast enough to counter an attack. But the person was also very important. The computers could keep the minds of the human operators “free to make only the necessary human judgements of battle - when and where to fight.’ (O’Neill, 1992, pp. 19-21) Whirlwind and its aftermath, Cape Cod and SAGE, inaugurated and defined a new technological paradigm, typically referred to as computerized real-time transaction processing systems (O’Neill, 1992, p. 22). This is a technology that facilitates workers in a cooperative effort in as much as the system provides them with a common field of work in the form of a data set or other type of digital representation (possibly coupled to facilities outside of the digital realm) and thus enables them to cooperate by changing the state of the data set in some strictly confined way: ‘On-line transaction processing systems were programmed to allow input to be entered at terminals, the central processor to do the calculation, and the output displayed back at the terminal in a relatively short period of time. These systems could be used, for example, to allow several users to make inquiries or update requests, which were handled as they arrived with only a short delay. The services that could be requested were restricted to those already pre-programmed into the system.’ (O’Neill, 1992, pp. 22-23) The SAGE system certainly facilitated cooperative work, in that state changes initiated by one worker propagated via the system to other workers, but from the 344 Cooperative Work and Coordinative Practices point of view of CSCW, one would not categorize it as a system that supported cooperative work. The operators were interdependent in their work by virtue of the system (radar stations, communication lines, tracking devices, handovers) and interacted in an orderly way regulated by the system but were strictly limited in what action and interaction they could undertake in and through the system. As O’Neill points out, ‘Although they could specify what information was displayed, the operators could not change the program or the situations’ (O’Neill, 1992, p. 21). Thus, when a US Air Force colonel at the time characterized the SAGE system as ‘a servomechanism spread over an area comparable to the American Continent’, he was quite right (Mindell, 2002, p. 313). The limits of this paradigm were evident from the very beginning. It was, for example, obvious that there would be situations where air-defense operators at Direction Centers would need to interact with colleagues, perhaps at another center, in ways different from ‘those already preprogrammed into the system’. Thus, to enable operators to handle contingencies, it was realized at a very early stage in the development of the system that operators would need to be able to interact outside of the functionalities afforded by the system. Thus, in the summer of 1953, a ‘“man-to-man” telephone intercommunication system’ was installed so as to enable operators in the Combat Center (of the Direction Center) and at some remote locations to simply talk to one another (Redmond and Smith, 2000, pp. 310 f.). That is, operators were not supported in enacting or developing coordinative practices by the computational system. The handling of contingencies had to be carried out outside of the computational system, as ordinary conversations among operators within each center or as telephone conversations among operators across centers. CSCW begins with the realization of these limitations in the Whirlwind paradigm. 2.5.2. The Whirlwind legacy The SAGE system had immediate and far-reaching impact on computing technologies. From the early 1960s, applications of the paradigm epitomized by SAGE such as computerized transaction processing systems for air traffic control, banking, manufacturing production planning and control, and inventory control began to upset the forms of work organization that had dominated administrative work settings for ages, either in the form of armies of ‘human computers’ organized on the basis of the de Prony principles or in the form of punched-card machine operators working in configurations similar to those of the cotton mills of Lancashire around 1830. It began with airline reservation systems. The first major civilian project to exploit the Whirlwind paradigm was the SABRE airline reservation system developed by IBM for American Airlines. It was, again, a critical situation in a cooperative work setting that motivated the development effort In 1954 a airline reservations office would come across somewhat like this portrayal of American Airline’s Chicago office: Chapter 11: Formation and fragmentation 345 ‘A large cross-hatched board dominates one wall, its spaces filled with cryptic notes. At rows of desks sit busy men and women who continually glance from thick reference books to the wall display while continuously talking on the telephone and filling out cards. One man sitting in the back of the room is using field glasses to examine a change that has just been made on the display board. Clerks and messengers carrying cards and sheets of paper hurry from files to automatic machines. The chatter of teletype and sound of card sorting equipment fills the air. As the departure date for a flight nears, inventory control reconciles the seating inventory with the card file of passenger name records. Unconfirmed passengers are contacted before a final passenger list is sent to the departure gate at the airport. Immediately prior to take off, no-shows are removed from the inventory file and a message sent to downline stations canceling their space.’ (McKenney, 1995, p. 97). An office like this would house about 60 reservation clerks and 40 ‘follow-up clerks’ and ‘board maintainers’. The logic of this scenery is as follows. In the 1930s American Airline had adopted a decentralized reservations system maintained at the particular flight’s departure point. This system, called ‘request-andreply’, involved significant coordination work, as it required agents to coordinate with the inventory control department (two messages) before confirming the reservation with a third message. In addition, passenger-specific data (name, telephone number, and itinerary) had to be recorded at the time of confirmation on a ‘passenger name record card’ and subsequently transmitted via telephone or teletype to the inventory control department. To reduce the cost of coordination, an amended system was introduced a decade later: until a flight was 80 percent sold out, agents were free to accept bookings and were only required to report actual sales, allowing passenger requests to be accommodated quickly. This reduced message volumes to half. At the same time, the inventory control department monitored sales, and when available seats decreased to a prescribed level, a ‘stopsale message’ was broadcast to all agents. To make this ‘sell and report’ system work required a buffer of seats. As McKenney observes, ‘These largely manual systems were time-consuming, requiring from several hours to several days to complete a reservation. Moreover, they were cumbersome and unreliable, plagued by lost names, reservations made but not confirmed with travelers, and inaccuracies in the passenger lists held at boarding gates, with the result that most business travelers returning home on a Friday would have their secretaries book at least two return flights.’ (McKenney, 1995, p. 98) At that point, although thousands of reservations were processed every day, the operation ran virtually without the use of machinery, apart from an electromechanical system at the inventory control department for maintaining an inventory of sold and available seats. In fact, as pointed out by Campbell-Kelly and Aspray, the reservations office ‘could have been run on almost identical lines in the 1890s’ (1996, p. 170). The reason for this was that existing data-processing technologies operated in batch-processing mode. The primary goal of punched-card tabulating machinery that dominated the administrative domain in large-scale business operations until this time was to reduce the cost of each transaction. This was done by completing each operation for all transaction records before the next operation on the records was begun. An individual transaction simply had to wait until an eco- 346 Cooperative Work and Coordinative Practices nomically viable batch of transactions had been accumulated. That is, the cost of each transaction was traded off against the lapse time of processing the transaction. According to Campbell-Kelly and Aspray, ‘the time taken for each transaction to get through the system was at least an hour, but more typically half a day’ (ibid.). As noted above, the new business computing technologies were designed and applied as little more than advanced punched-card tabulators. What these technologies offered was not another mode of operation but batch-processing at a higher level of automation. With punched-card tabulating machinery human involvement was required in the transition from one batch process to the next: cards had to be picked up and carried from, say, sorting to tabulating to re-sorting. But with computers equipped with tape stations this vestige of human intervention could be eliminated. When the transaction represented by the punched cards had been ‘read’ by the system’s card reader, the worker could stand back and supervise the entire process. For most businesses, batch processing was a cost-effective mode of operation; it was how accounting offices updated accounts, banks processed checks, insurance companies issued policies, and utility companies invoiced clients. But the airline industry, by contrast, had ‘instantaneous processing needs’, and trading transaction cost off against lapse time was not an option. Hence the overwhelmingly manual character of the entire reservations operation by 1954. But in the 1950s the volume of traffic and the number of flights increased. Thus ‘the ability to process passenger reservations assumed increased importance. Given the highly perishable nature of the airlines’ product, reservation personnel needed to have rapid access to seat availability, be able to register customer purchases instantly, and in the event of a cancellation quickly recognize that a valuable item of stock had been returned to inventory and was available for resale.’ (McKenney, 1995, pp. 98 f.). Consequently, operating the reservations office as a manufactory was no longer viable. The ‘boards became more crowded’ and the clerks ‘had to sit farther and farther away from them’. So, by 1953 ‘American Airlines had reached a crisis.’ (Campbell-Kelly and Aspray, 1996, pp. 170 f.). Having learned of the real-time transaction processing technology that was then under rapid development, American Airlines decided to undertake a thorough transformation of its reservation system based on this paradigm. The new system was to be developed by IBM (where Perry Crawford had been employed in 1952 to develop real-time applications). The project, later dubbed SABRE, was formally established in 1957 and was the ‘at the time easily the largest civilian computerization task ever undertaken’. The system was implemented in the early 1960s on IBM 7090 computers, which were for all practical purposes solid-state versions of the Q-7. The final system, which was fully operational in 1964, connected some 1,100 agents using desktop terminals across the US, who had access to ‘a constantly updated and instantly accessible passenger name record’ containing various information about the passenger, including telephone contacts, hotel and automobile reservations, etc. (Campbell-Kelly and Aspray, 1996, pp. 172-174; Campbell-Kelly, 2003). As in the SAGE system, the SABRE system facilitated Chapter 11: Formation and fragmentation 347 cooperative work: one worker’s action on computational objects (or structured data sets) represented in the system in some form, e.g., his or her entering a flight reservation (name, date, flight number, etc.), is stored (and perhaps transmitted to another center) to allow other workers (perhaps elsewhere) to perceive and act on this information. In other words, the system mediates the propagation of changes to the state of the common field of work of the many workers involved in the large-scale cooperative effort of handling flight reservations. In this paradigm the computational system does not support articulation work. What characterizes these systems is real-time online transaction processing. ‘Transaction processing allows a person to interact with the system in ways that are limited and predetermined’ (O’Neill, 1992, p. 2). Whatever might be required to handle a given task in addition to and beyond what the transaction contains, e.g., sorting out ambiguities, mistakes, etc., is done outside of the system, through some other medium, e.g., by telephone. In short and rather schematically, realtime online transaction processing systems facilitate cooperative work but not articulation work. From the vantage point of CSCW, there are many reasons to take note of the Whirlwind project and its aftermath. As already described, Whirlwind provided the paradigm for a computing technology in which the interdependent activities of multiple of workers (hundreds, thousands even) are facilitated and mediated by the system. More than that: Whirlwind played a crucial and unique role in the development of computing technology beyond the paradigm of massive calculation: digital communication networks, real-time computing, interactive computing, direct manipulation, etc. The conceptual foundation of much of modern computing was developed in the course of this effort. 2.5.3. Interactive computing Interactive computing grew out of technologies and practices far away from scientific and administrative calculation. It grew out of technologies of real-time control of external processes. The stored-program digital computer provided the very possibility of extending and transforming the technologies of real-time control of external processes, but it is no accident that Whirlwind was developed by MIT’s Servomechanism Laboratory. The interactive computing paradigm as developed in the course of Whirlwind and so on extended and transformed technologies of gunnery control that in turn build on knowledge of real-time processes such as anti-aircraft control, bomb aiming, automatic aircraft stabilizers, etc. However, the Whirlwind legacy is not simply a faint pattern constructed in retrospect by the historian. The continuity was real to actors at the time. First of all, there was a direct personal continuity: ‘The people who worked on the two projects gained an appreciation for how people could interact with a computer. This appreciation helped the development and maturing of interactive computing, linked with time-sharing and networking’ (O’Neill, 1992, p. 11). More than that, the Whirlwind project and the subsequent development of the SAGE air defense system were research and development projects on a huge scale. The 348 Cooperative Work and Coordinative Practices overall cost of the SAGE system exceeded the cost of the Manhattan project by a wide margin. The project therefore had tremendous impact on American computer science and engineering communities. In fact, half of the trained programming labor force of the US at the time (1959) were occupied with the development of the software for the SAGE system (Campbell-Kelly, 2003, p. 39). (Cf. also Baum, 1981). Secondly, when the Whirlwind project was finished in 1953 and many researchers from the Servomechanisms Lab were relocated to work on Cape Cod and SAGE at a new laboratory (the Lincoln Laboratory), the original Whirlwind prototype was not scrapped but remained available on the MIT campus. Furthermore, at the Lincoln Laboratory researchers developed transistorized experimental computers based on the Whirlwind architecture. Named TX-0 and TX-2, they ‘provided quick response and a variety of peripheral input/output equipment which allowed these machines to be used interactively’. These computers were also handed over to MIT campus in 1958 and were used intensively by graduate students and researchers: ‘The attitude of the people using the Whirlwind computer and these test computers was important in the establishment of a “culture” of interactive computing in which computers were to be partners with people in creative thinking’ (O’Neill, 1992, p. 24). For example, Fernando Corbató, who later played a central role in the development of time-sharing, recalls that ‘many of us [at MIT] had cut our teeth on Whirlwind. Whirlwind was a machine that was like a big personal computer, in some ways, although there was a certain amount of efficiency batching and things. We had displays on them. We had typewriters, and one kind of knew what it meant to interact with a computer, and one still remembered.’ (Corbató, 1990, p. 14). In fact, the interactive computing ‘culture’ O’Neill refers to was so ingrained among Whirlwind programmers that they were not particularly interested in time-sharing when this movement got under way but were rather, in Corbató’s words, preoccupied with ‘creating in some sense a humongous personal computer (laugh). This included people like Ivan Sutherland, for example, who did the Sketchpad using TX-2. That was a direct result of being able to have it all to himself, as though it were a big personal computer.’ (Corbató, 1990, p. 15). Thus the protracted and costly research effort that went into building the Whirlwind and its offspring provided the basic principles and techniques of the interactive computing paradigm. Later developments extended and refined the notion of ‘interacting’ with a computer in ‘real time’ far beyond the initial and quite narrow constraints of the concept of real-time transaction processing and also beyond other Whirlwind technologies such as CRT displays, and so on. Major steps were taken already in the 1960s, in laboratories using Whirlwind-type computers such as the TX-2. Ivan Sutherland’s Sketchpad (1963) is probably the first documented implementation of modern computer graphics technologies. A series of major advances resulted from the tenacious research effort of Douglas Engelbart who, of course, was well aware of the interactive computing paradigm when this effort began (Engelbart, 1962, §IIIA5; cf. also Bardini, 2000, Chapter 11: Formation and fragmentation 349 passim). It is well known that the first experimental versions of the computer mouse was developed by Engelbart’s laboratory at the Stanford Research Institute (W. K. English, et al., 1967), but a significantly more important step was taken in 1968 when Engelbart at an AFIPS conference demonstrated that many of the technologies that we now consider essential components of interactive computing (direct manipulation, bit-mapped displays, mouse, message handling, etc.) could be realized in an integrated fashion on the basis of an architecture providing unrestrained access to multiple application programs (Engelbart and English, 1968). Some of the crucial points on the further development trajectory of this technology (or web of technologies) include the Xerox Alto from 1973 (Thacker, et al., 1979; Lampson, 1988; Thacker, 1988) and the Xerox 8010 ‘Star’ from 1981 (D. C. Smith, et al., 1982a; D. C. Smith, et al., 1982b; J. Johnson, et al., 1989). Interactive computing, as we know it, finally became a practical reality with the release of the Apple Macintosh in 1984. 2.5.4. The arrested growth of interactive computing Whirlwind went online in 1951, development of the Cape Cod system was finished around 1956, and the SAGE system went operational around 1960. Still, it took a generation for the technologies of interactive computing to become a practical reality for workers outside of a small population of computer scientists and workers in certain time-critical work settings such as air defense and airline reservation. What impeded the further development of interactive computing, beyond on-line transaction processing? The answer is not only that the development of Whirlwind was funded at a generous level that hardly has been seen since then, but also that the basic principles of interactive computing could not be disseminated beyond the few areas where their application was economically sustainable. For computing equipment was enormously expensive until the development of integrated circuits in the 1960s and especially microprocessors in the 1970s made mass production of computers economically viable.96 Semiconductor technology was very long in the making, and involved advances in fundamental theories of physics as well as the development of mass markets for electronics products. Researchers in the area of solid state physics had investigated the electrical properties semiconductors since the 1830s (although not by that name). What caught their attention was the fact that the electrical resistance of substances such as silver sulphide increased with increasing temperature, contrary to the behavior of ordinary conductors. In 1874 it was discovered (by Ferdi96 Excellent general accounts of the history of semiconductor technology can be found in the studies by Braun and Macdonald (1978) and by Orton (2009). The development of the transistor and integrated circuit technologies is recounted by the historians Riordan and Hoddeson (1997). Based on interviews with key actors, Reid (2001) gives a readable journalistic account of the semiconductor history from the beginnings to the invention of the microprocessor but his account is unfortunately marred by American hero-worship. — On the other hand, Bo Lojek (2007) offers a detailed and candid story of this development, reminding us that the history of modern technology is somewhat distorted by the picture of technology development as ‘a systematic effort of exceptional leadership’ carefully nurtured by corporate PR departments and patent lawyers and that, based on his own experience as an engineer in the semiconductor industry, ‘the company establishment was frequently one of the biggest, if not the biggest, obstacle’. 350 Cooperative Work and Coordinative Practices nand Braun) that a metal wire contacting a crystal of lead sulphide (another semiconductor) would conduct electricity in one direction only. The rectifying effect, as it was called, was used in very simple radio receivers (‘crystal radio’) in the first decades of the twentieth century, but the effect could not be not explained until the development of quantum mechanics in the 1920s and 1930s provided the theoretical framework for beginning to model the behavior of electrons in solids. But it was not until the end of the 1930s that the rectification effect was explained (by Mott, Schottky, and Davydov) and the theoretical basis for the invention of the transistor was laid, and even then an enormous systematic effort — e.g., charting the properties of semiconductors such as germanium and silicon in the course of developing radar technology during World War II — was required to arrive at the point when the transistor effect was actually discovered (by Bardeen, Shockley, and Brattain at Bell Labs on 23 December 1947). Still, as Braun and Macdonald put it in their history of semiconductor technologies, even then ‘a vast amount of development work remained to be done if the transistor was ever to become a technological achievement rather than just a scientific curiosity’. ‘This is a case of a very large gulf of ignorance separating the early ideas from any possibility of realisation. Much knowledge had to be accumulated before the original ideas could be modified so as to create practical devices’ (Braun and Macdonald, 1978, pp. 24 f., 47). So, although transistors were in commercial production from 1951, the technology did not affect the computer industry for more than a decade. The production of transistors posed ‘staggering problems’. Vital parameters such as conductivity of a semiconductor crystal depend critically on the nature, amount, and distribution of non-native atoms in the crystal lattice (at degrees of accuracy of 1 of 108), which made semiconductor production an exceedingly difficult art to master. In the early 1950s the price for one transistor was about $20 while the price for a thermionic valve (or vacuum tube) was about $1; by the end of 1953, when the annual production reached 1 million transistors, the price was about $8. This initially excluded transistors from supplanting thermionic valves in most civilian applications. One of the first applications was in hearing aids, which began around 1953, and by the end of 1954 the portable radio was launched, but in general the technology only developed slowly, largely driven by military applications and the US space program. Anyway, at this time the transistor was not nearly as reliable as the thermionic valve. For several years, the choice material for transistors was germanium, which is far more pliant in the production process but also highly sensitive to variations in temperature and not very suitable for making the switches that make up digital circuitry in the first place. Silicon, on the other hand, would afford a far more reliable transistor but was also very difficult to work with. Silicon transistors therefore only became an economic reality for the computer industry with the invention of integrated circuits in the early 1960s. Consequently, the computer industry stayed with the thermionic valve until around 1960 when printed circuits provided a way to apply transistor technology in large-scale production of computers such as the IBM 1401. Chapter 11: Formation and fragmentation 351 In addition to the ‘staggering problems’ encountered in the production of transistors, the application of the technology in computing was impeded by another problem, an impasse called the ‘tyranny of numbers’. Circuits consisted of discrete components (transistors, resistors, capacitors, etc.) that had to be handled individually. In the words of Jack Morton, a manager semiconductor research at Bell Labs, ‘Each element must be made, tested, packed, shipped, unpacked, retested, and interconnected one-at-a-time to produce the whole system’. To engineers it was evident that this meant that the failure rate of circuitry would increase exponentially with increased complexity. Thus, so long as large systems had to be built by connecting ‘individual discrete components’, this posed a ‘numbers barrier for future advances’ (quoted in Reid, 2001, p. 16). No matter how reliable the individual components were, the circuits were only as reliable as the connections, and connections were generally manually wired: ‘The more complex the system, the more interconnections were needed and the greater the chance of failure through this cause’ (Braun and Macdonald, 1978, p. 99). The integrated circuit — an entire circuit made out of one monolithic crystal — was developed to break this barrier. The idea was conceived and developed by several researchers interpedently of each other (Jack Kilby at Texas Instruments and Robert Noyce at Fairchild Semiconductors). As Kilby expressed his idea in his notebook on 24 July 1958: ‘The following circuit elements could be made on a single slice: resistors, capacitor, distributed capacitor, transistor’ (quoted in Reid, 2001, p. 77). The tedious and error-prone manual process of testing and assembling circuits from discrete components could be eliminated. However, it was not the application of integrated circuits in computers that primarily motivated and drove the development of this technology but more or less frivolous applications such as digital wristwatches and pocket calculators. But as the integrated circuit technology matured and prices fell, small computer manufacturers such as Digital Equipment and Data General adopted the integrated circuit as the basis of compact ‘minicomputers’ that could be sold for less than $100,000. At this stage, interactive computing in the form of engineering workstations slowly began to become a reality beyond a few exceptionally well-equipped computing research laboratories. However, it took the microprocessor to make interactive computing an economic reality. It is worth noting that the microprocessor was not developed for the computer industry at all, but was conceived as a means of reducing the cost and time required to design circuitry for appliances such as electronic calculators. One could say that the ‘numbers barrier’ had been replaced by a ‘circuit designer barrier’. In 1969, Intel, then a new company, landed a contract to develop a set of calculator chips for Hayakawa Electric Co. Tasked to work on the project, Marcian E. Hoff, suggested to make a general-purpose computer chip instead of a set of hard-wired calculators. The idea was subsequently developed by Stan Mazor and Federico Faggin and others and the first ‘micro-programmable computer on a chip’, the Intel 4004, was eventually released at the end of 1971, about 35 years after the discovery of the transistor and the demonstration of the first stored-program digital 352 Cooperative Work and Coordinative Practices computer and almost 30 years after the conception of interactive computing was first realized in Whirlwind. — Intel 4004 was followed by the 8008 (8-bit wordlength processor) chip, presented in 1973. Parallel developments occurred at Texas Instruments. By the mid-seventies the microprocessor was an established and accepted technology complete with computer-aided design environments: ‘In no more than five years a whole industry had changed its emphasis from the manufacture of dedicated integrated circuits [for pocket calculators, etc.], to the manufacture of what in effect were very small computers’ (Braun and Macdonald, 1978, p. 110). Where the 4-bit Intel 4004 of 1971 contained 2,300 devices (transistors, etc.), the 32-bit processors that began to be produced about ten years later integrated about 200,000 transistors and were comparable with mainframe computers in performance (ibid., pp. 108-112). As pointed out by Braun and Macdonald (ibid.), underlying this remarkable leap was a ‘favorable constellation’ of factors in the semiconductor industry. Firstly, the invention of electronic calculators and their success had created a mass market for integrated circuits. Advances in solid-state physics in general and semiconductor manufacturing technologies in particular (e.g., clean-room fabrication of ‘metal–oxide–semiconductors’ or MOS by means of photolithography) made it possible to produce integrated circuits with increasing densities and low power consumption. Secondly, related technical developments in high-density MOS technology made it possible to produce semiconductor memory chips. Thirdly, minicomputer technology had matured to the stage that their hardware architectures could serve as a model for microprocessor design. And so on. The microprocessor, initially developed for calculators, was shifted laterally and formed a critical component technology in the personal computer that was then created. In the words of Douglas Bell, the designer of many of Digital’s PDP computers, in his keynote address to the ACM conference on the History of Personal Workstations, ‘The microprocessor, memory, and mass storage technology appearing in 1975 lead directly to the personal computer industry’. Improvements in these technologies were ‘the sole enabling determinants of progress in computing’ (1988, pp. 9 f.). That is, when Engelbart undertook his experimental work on the NLS in the second part of the 1960s, integrated circuits were in their infancy. The computer platform that was available to his Augmentation Research Center for its experiments consisted of a couple of CDC minicomputers, but in 1967, the laboratory, thanks to a grant from ARPA, was able to acquire its first time-sharing computer; it cost more than $500,000 (Bardini, 2000, pp. 123, 251). The Xerox Alto, released only seven years later, in 1974, had a performance comparable to that of a minicomputer but the cost had dwindled to about $12,000. This was still well above what was commercially viable, but the Alto was anyway primarily developed as an experimental platform. Eventually 1,500 machines were built and used by researchers at Xerox and at a few universities (Thacker, 1988). ‘Alto served as a valuable prototype for Star. Over a thousand Altos were eventually built, and Alto users have had several thousand work-years of experience with them over a period of eight Chapter 11: Formation and fragmentation 353 years, making Alto perhaps the largest prototyping effort in history. There were dozens of experimental programs written for the Alto by members of the Xerox Palo Alto Research Center. Without the creative ideas of the authors of those systems, Star in its present form would have been impossible.’ (D. C. Smith, et al., 1982b, p. 527 f.). In other words, due to the drastically reduced costs of computing power due to microprocessor technology, an experimental platform for interactive computing research had become available. And then, eventually, one day in 1975 Apple cofounder Stephen Wozniak could walk in and buy a MOS Technology 6502 microprocessor for $20 (Freiberger and Swaine, 2000, p. 264), and less than ten years later Apple could release the first Macintosh at a price about $1,200. The technology of interactive computing was, so to speak, dormant for a decade or more, until Moore’s law propelled the technology forward (cf. alsoGrudin, 2006a, b). The long gestation period was not due to theoretical shortcomings. For the technologies of interactive computing were never derived from any preexisting theoretical knowledge. In fact, the technologies of interactive computing were developed by computer technicians to satisfy requirements they themselves had formulated on the basis of principles and concepts known from their own daily work practices. On one hand, in the design of the Alto and the Star the technicians built on and generalized concepts that were deeply familiar to any Western adult working in an office environment. In their own words, ‘the Star world is organized in terms of objects that have properties and upon which actions are performed. A few examples of objects in Star are text characters, text paragraphs, graphic lines, graphic illustrations, mathematical summation signs, mathematical formulas, and icons’ (D. C. Smith, et al., 1982b, p. 523). On the other hand, in interpreting these everyday concepts technically the designers applied the conceptual apparatus of objected-oriented programming (objects, classes of objects, properties, messages) that had been developed by Kristen Nygaard and Ole-Johan Dahl as a technique to obtain ‘quasi-parallel processing’ (Dahl and Nygaard, 1966; cf. also Nygaard and Dahl, 1978) and had later been further developed in the 1960s by Alan Kay and others in the form of Smalltalk: ‘Every object has properties. Properties of text characters include type style, size, face, and posture (e.g., bold, italic). Properties of paragraphs include indentation, leading, and alignment. Properties of graphic lines include thickness and structure (e.g., solid, dashed, dotted). Properties of document icons include name, size, creator, and creation date. So the properties of an object depend on the type of the object.’ (D. C. Smith, et al., 1982b, p. 523) Similarly, the technicians could build on ‘fundamental computer science concepts’ concerning the manipulation of data structures in order to provide application-independent or ‘generic’ commands that would give a user the ability to master multiple applications and to ‘move’ data between applications: ‘Star has a few commands that can be used throughout the system: MOVE, COPY, DELETE, SHOW PROPERTIES, COPY PROPERTIES, AGAIN, UNDO, and HELP. Each performs the same way regardless of the type of object selected. Thus we call them generic commands. […] These commands are more basic than the ones in other computer systems. They strip away extra- 354 Cooperative Work and Coordinative Practices neous application specific semantics to get at the underlying principles. Star’s generic commands are derived from fundamental computer science concepts because they also underlie operations in programming languages. For example, program manipulation of data structures involves moving or copying values from one data structure to another. Since Star’s generic commands embody fundamental underlying concepts, they are widely applicable. Each command fills a host of needs. Few commands are required. This simplicity is desirable in itself, but it has another subtle advantage: it makes it easy for users to form a model of the system.’ (D. C. Smith, et al., 1982b, p. 525). The important point here is that the design concepts reflected the technicians’ own practical experience, in that they could generalize the concepts of their own work (typographical primitives such as characters, paragraphs, type styles, etc.) and from similar concepts already generalized in computer science (MOVE, COPY, DELETE, etc.). No ethnographic fieldwork was required to develop an understanding of concepts like character, paragraph, line, and illustration or move, copy, and paste. For a long stretch, then, interactive computing technologies could be further developed and refined in a manner very similar to the way in which email was developed: by technicians building tools for their own use, based on their daily practices and in a rather intuitive iterative design and evaluation process. Thus, for example, when Steve Wozniak was designing the Apple II, he ‘was just trying to make a great computer for himself and impress his friends at the Homebrew Computer Club. […] Most of the early Apple employees were their own ideal costumers.’ The members of the Macintosh design team were similarly motivated: ‘we were our own ideal customers, designing something that we wanted for ourselves more than anything else’ (Hertzfeld, 2005, pp. xvii f.). They knew the concept of interactive computing from Engelbart’s work (who had been building on the Whirlwind experience); they knew from their own lives what they would require of its design; and they had the advanced technical skills to realize their ideas. That is, the technologies of interactive computing were initially developed in the course of a deliberate design effort (in the Whirlwind and Cape Cod projects), drawing upon on principles of ‘man-machine systems’ based on experiences from servomechanisms. After that, the technologies of interactive computing were subjected to two decades of almost ‘arrested growth’. However, the technology of microprocessors, mass-produced CPUs provided a burgeoning platform of development on which the computer scientists at SRI, at Xerox PARC, at Apple, and elsewhere could extend, elaborate, and refine the principles of interactive computing on the basis of their own practical experiences The first CHI conference convened in 1982, that is, one year after the release of the Xerox Star and two years before the release of the first Macintosh. That is, as an institutionalized research field, HCI only emerged after the technology had matured and was on the way out of the laboratories. This has caused all kinds of soul searching. Chapter 11: Formation and fragmentation 355 As shown earlier, John Carroll noted that ‘some of the most seminal and momentous user interface design work of the last 25 years’ such as Sutherland’s Sketchpad and Engelbart’s oN Line System (NLS) ‘made no explicit use of psychology at all’ (Carroll, 1991, p. 1). Carroll’s observation is fully vindicated by the original documents, although it should be mentioned that psychologists at RAND and the SDC such as Allen Newell were involved in the development of the SAGE system software already from the 1950s, tasked with testing the software modules that were to support operators at the air defense direction centers (Chapman, et al., 1959; Baum, 1981). That is, psychology did contribute to the development of interactive computing but the contribution was made almost a decade prior to Sutherland and Engelbart’s work and had become incorporated in the interactive computing paradigm as represented by real-time transaction processing technology that formed the technological baseline for their work.97 As already noted, Carroll’s observations also apply to the development of the Xerox Alto and the Xerox Star. But again a little nuance is required, in as much as human factors researchers were involved in ‘human factors testing’ of the design of the Star. Some of that work was reported at the first CHI conference by Bewley et al. (1983). However, in the same paper Bewley et al. claim that ‘Recognizing that design of the Star user interface was a major undertaking, the design team approached it using several principles, derived from cognitive psychology’ (p. 72). The authors provide no justification for this claim, nor is it supported by the available contemporary documentation concerning the ‘Star’ design process. In fact, the very general features they cite were already present in the design of the Alto that was designed ten years earlier. The claim thus has the appearance of a rationalization. On the other hand, just for the record, the development of the mouse did draw upon principles from experimental psychology, in particular ‘Fitts’ law’ (1954) concerning the capacity of the human motor system (cf. D. C. Smith, et al., 1982b). Anyway, and as a matter of fact, the principles of ‘direct manipulation’ were only articulated and systematized a decade after the release of the Xerox Alto (Shneiderman (1983) and Hutchins, Hollan, and Norman (1986)). HCI research developed as a post festum systematization effort devoted to understanding the principles of the interactive computing techniques that had emerged over three decades, from the Whirlwind to the Alto, the Star, and the Macintosh. 2.5.5. Interactive computing and the cybernetic notion of ‘human-computer system’ Interactive computing had its intellectual roots in the real-time transaction processing technologies that in turn emerged from control-engineering conceptions of human control of processes in ‘real time’ (gun control, bomb aiming, servo steering). Generating research traditions such as ‘man-machine systems’ research, this 97 Stuart Card remarks cannily that what Newell was engaged in at RAND and SDC today would be called ‘Computer Support for Cooperative Work’ (1996, p. 259), a claim not without merit. 356 Cooperative Work and Coordinative Practices conception of machinery has left a rich and quite influential legacy: a conceptual framework that has been expressed in different terms over time — from the original notions of ‘man-machine system’ (Sutherland, 1963), ‘man-computer symbiosis’ (Licklider, 1960), and ‘human intellect’ ‘augmentation system’ (Engelbart, 1962) to contemporary notions of ‘human-computer interaction’ and ‘interaction design’ — but the underlying conception is the same. What unites all these different notions is the concept of unity of control in the man-machine system dyad: the notional steersman directing the dyad towards his or her ‘goal’. It is thus, invariably, presumed that control is vested in an individual or in a ‘team’ or ‘community’ acting in such as way that unity of control can be presumed as a given. The argument can be made that the basic concept of HCI — the humancomputer dyad — still reflects this outlook and that the defining conceptual axis of HCI research has remained that of the human-machine dyad. It goes without saying that communication and interaction among people is not excluded from this conception, but the conception is incommensurable with conceptions of the essentially distributed nature of cooperative work and of the sociality of human competencies. This becomes evident when attempts are made to address technical requirements of cooperative work under the aegis of this conception, such as, for example, when Engelbart and Lehtman in 1988 outlined their vision of a ‘community handbook’ as ‘a monolithic whole’, ‘a uniform, complete, consistent, upto-date integration of the special knowledge representing the current status of the community’ (Engelbart and Lehtman, 1988), the essentially distributed nature of such a construct was entirely absent from the conception. The very same conception underlies the ‘groupware’ research paradigm: computing technologies that facilitate the exchange of messages and files and thereby ‘collaboration’ or ‘group work’ or ‘shared knowledge’ among dispersed individuals working together ‘at a distance’ as a ‘group’ or ‘team’ towards a ‘collaborative goal’ (cf., for example, G. M. Olson and Olson, 2003). The problem with such notions is that the issue of the distributed character of cooperative work — the issue of the distributed control of coordination — is reduced to issues of geographical dispersion: the issues of heterogeneity of practices, incongruence and incommensurability of conceptual schemes, etc. are not reflected and cannot be integrated in the cybernetic ‘human-machine system’ conception. 2.6. Facilitation of articulation work: Computer-mediated communications The notion of computer-mediated communications began with the notion of ‘time-sharing’ operating systems that matured around 1960. Since computer systems at the time were excessively expensive (printed circuit boards and core memory was only then becoming standard technologies), it was mandatory that computer systems were operating close to full capacity. Consequently, most of the few computers that were around were running in a batch-processing mode, one job after another on a ‘first-in, first-served’ basis, or, as it was aptly expressed by J. C. R. Licklider, the ‘conventional computer-center mode of operation’ was ‘pat- Chapter 11: Formation and fragmentation 357 terned after the neighborhood dry cleaner (“in by ten, out by five”)’ (Licklider and Clark, 1962, p. 114). This economic regime could be tolerated for administrative purposes such as payroll and invoice processing that, since punched-card tabulating had been adopted earlier in the century, was organized on batch-processing principles anyway. But in most work settings, this regime effectively precluded all those applications that require high-frequency or real-time interactions between user and the digital representations. In particular, the ‘in by ten, out by five’ regime made programming, especially debugging, a deadening affair. This gave ordinary computer technicians a strong motive for devising alternative modes of operation. As described by O’Neill, researchers at MIT, who had ‘cut their teeth’ on Whirlwind and had experienced interactive computing first-hand, ‘were unwilling to accept the batch method of computer use for their work’ and ‘sought ways to continue using computers interactively’. However, ‘it would be too costly to provide each user with his or her own computer’ (O’Neill, 1992, p. 44). So, around 1960 the idea of letting a central computer system service several users ‘simultaneously’ was hatched. In the words of John McCarthy, one of the fathers of the idea, the solution was an operating system that would give ‘each user continuous access to the machine’ and permit each user ‘to behave as though he were in sole control of a computer’ (McCarthy, 1983). The first running operating system of this kind seems to have been the Compatible Time-Sharing System or CTSS. It was built at MIT by a team headed by Fernando Corbató and was first demonstrated in 1961 (Corbató, et al., 1962). The various users were connected to the ‘host’ computer via terminals and each would have access to the computing power of the ‘host’ as if he or she was the only user. Now, the users of the first of these systems were typically engaged in cooperative work. Some were engaged in developing operating systems or other largescale software projects and were, as a vital aspect of this, engaged in various forms of discourse with colleagues within the same project teams and research institutions, that is, with colleagues already connected to the central computer system. Likewise, software technicians would need to coordinate with system operators about possibly lost files to be retrieved, about eagerly-awaited print jobs in the queue, etc. The time-sharing operating system they were building or using provided a potential solution to this need, and the idea of using the system to transfer text messages from one worker to another did not require excessive technical imagination. As one of the designers of one of the first email systems recalls: ‘[CTSS] allowed multiple users to log into the [IBM] 7094 from remote dial-in terminals[] and to store files online on disk. This new ability encouraged users to share information in new ways. When CTSS users wanted to pass messages to each other, they sometimes created files with names like TO TOM and put them in "common file" directories, e.g. M1416 CMFL03. The recipient could log into CTSS later, from any terminal, and look for the file, and print it out if it was there.’ (Van Vleck, 2001) A proper mail program, ‘a general facility that let any user send text messages to any other, with any content’ was written for CTSS by Tom Van Vleck and Noel 358 Cooperative Work and Coordinative Practices Morris in the summer of 1965 (ibid.). It allowed one programmer to send a message to other individual programmers, provided one knew the project they worked on, or to everybody on the same project. The message was not strictly speaking ‘sent’; it was appended to a file called MAILBOX in the recipient’s home directory. The same year Van Vleck and Morris also devised a program (.SAVED) ‘that allowed users to send lines of text to other logged-in users’, that is, a primitive form of ‘instant messaging’ (ibid.). The scope of the exchange of messages with these and similar programs was limited by the boundary of the hierarchy comprising the local central computer system and the terminals connected to it. Messages could not travel beyond the event horizon of this black hole. This world of isolated systems dissolved with the development of network computing. The motivation driving the development was (again) not to develop facilities for human interaction, not to mention cooperative work, but to utilize scarce resources in a more economical way. For Licklider, who also initially headed the development of ARPANET, the motivation for the network was to reduce ‘the cost of the gigantic memories and the sophisticated programs’. When connected to a network, the cost of such shared resources could be ‘divided by the number of users’ (Licklider, 1960, p. 8). This vision was spelled out by Lawrence Roberts, who took over as manager of the development of ARPANET in 1966. Noting that the motivation for past at- tempts at computer networks had been ‘either load sharing or interpersonal message handling’, he pointed to ‘three other more important reasons for computer networks […], at least with respect to scientific computer applications’, namely: ‘Data sharing: The program is sent to a remote computer where a large data base exists.’ ‘Program sharing: Data is sent to a program located at a remote computer and the answer is returned.’ ‘Remote Service: Just a query need be sent if both the program and the data exist at a remote location. This will probably be the most common mode of operation until communication costs come down [since this] category includes most of the advantages of program and data sharing but requires less information to be transmitted between the computers.’ (Roberts, 1967, p. 1). On the other hand, electronic mail was explicitly ruled out as of no significance: ‘Message Service: In addition to computational network activities, a network can be used to handle interpersonal message transmissions. This type of service can also be used for educational services and conference activities. However, it is not an important motivation for a network of scientific computers’. (Roberts, 1967, p. 1). These motivations for developing the ARPANET were upheld and confirmed as the net began to be implemented: ‘The goal of the computer network is for each computer to make every local resource available to any computer in the net in such a way that any program available to local users can be used remotely without degradation. That is, any program should be able to call on the resources of other computers much as it would call a subroutine. The resources which can be shared in this way include software and data, as well as hardware. Within a local community, time-sharing systems Chapter 11: Formation and fragmentation 359 already permit the sharing of software resources. An effective network would eliminate the size and distance limitations on such communities.’ (Roberts and Wessler, 1970, p. 543). Thus, as summarized by Ian Hardy, in his very informative history of the origins of network email, the primary motive was economic. ‘ARPANET planners never considered email a viable network application. [They] focused on building a network for sharing the kinds of technical resources they believed computer researchers on interactive systems would find most useful for their work: programming libraries, research data, remote procedure calls, and unique software packages available only on specific systems.’ (Hardy, 1996, p. 6). After pioneering work on the underlying packet-switching architecture and protocols, the experimental ARPANET was launched in 1969, connecting measly four nodes.98 In the summer of 1971, when the network had expanded to fifteen nodes, a programmer named Ray Tomlinson at BBN (Bolt, Beranek, and Newman), devised a program for sending email over the network. He recalls that, while he was making improvements to a single-host email program (SNDMSG) for a new time-sharing operating system (TENEX) for the PDP-10 minicomputer, ‘the idea occurred to [him]’ to combine SNDMSG with en experimental file-transfer protocol (CPYNET) to enable it to send a message across the network, from one host to another, and append it to the recipient’s MAILBOX file. For that purpose he also devised the address scheme NAME@HOST that has become standard (Hardy, 1996; Tomlinson, 2001). ‘The first message was sent between two machines that were literally side by side. The only physical connection they had (aside from the floor they sat on) was through the ARPANET’, that is, through the local Interface Message Processor (IMP) that handled packet switching. To test the program, he sent a number of test messages to himself from one machine to the other. When he was satisfied that the program seemed to work, he sent a message to the rest of his group explaining how to send messages over the network: ‘The first use of network email announced its own existence’ (Tomlinson, 2001). The program was subsequently made available to other sites on the net that used TENEX on PDP-10s and was soon adapted the IBM 360 and other computers (Salus, 1995, p. 95). An instant success within the tiny world of ARPANET programmers, this very first network email program triggered a chain reaction of innovation that within less than a couple of years resulted in the email designs we use today: a list of available messages indexed by subject and date, a uniform interface to the handling of sent and received mail, forwarding, reply, etc. — all as a result of programmers’ improving on a tool they used themselves. In 1977, an official ARPANET standard for electronic mail was adopted (Crocker, et al., 1977). The history of network email after that is well known. The technology migrated beyond the small community of technicians engaged in building computer 98 Much of the pioneering work on packet switching was done by Donald Davies and his colleagues at the British National Physical Laboratory (Davies, 1965, 1966). The NPL network, which was also launched in 1969, was built with rather similar objectives in mind (On NPLNet, cf. also Abbate, 1999). 360 Cooperative Work and Coordinative Practices networks to computer research in general and from there to the world of science and eventually to the world at large. What is particularly remarkable in this story, and what also surprised those involved when they began to reflect on the experience, was ‘the unplanned, unanticipated and unsupported nature of its birth and early growth. It just happened, and its early history has seemed more like the discovery of a natural phenomenon than the deliberate development of new technology’ (Myer and Dodds, 1976, p. 145). And at a meeting in January 1979, convened to discuss the ‘the state of computer mail in the ARPA community and to reach some conclusions to guide the further development of computer mail systems’, it was ‘noted’ as a fact ‘that most of the mail systems were not formal projects (in the sense of explicitly sponsored research), but things that “just happened”’ (Postel, 1982, p. 2). In the same vein, the official ARPANET Completion Report notes that ‘The largest single surprise of the ARPANET program has been the incredible popularity and success of network mail’ (Heart, et al., 1978, p. III-110). Network email was not only ‘unplanned’ and ‘unanticipated’; it was ‘mostly unsupported’ (Heart, et al., 1977, p. III-67; Abbate, 1999, p. 109), for, as noted already, the objective of the ARPANET was resource sharing. Not only had Lawrence Roberts not included network mail in the original plans for the network, he had excluded it as ‘not an important motivation for a network of scientific computers’ (Roberts, 1967, p. 1). However, as Janet Abbate points out in her history of the emergence of the Internet, the aimed-for resource sharing failed to materialize. By 1971, when the original fifteen sites for which the net had been built were all connected, ‘most of the sites were only minimally involved in resource sharing’ (Abbate, 1999, p. 78). She notes that the ‘hope that the ARPANET could substitute for local computer resources was in most cases not fulfilled’, and adds that, as the 1970s progressed, ‘the demand for remote resources actually fell’, simply because minicomputers ‘were much less expensive than paying for time on a large computer’ (ibid., pp. 104 f.). The budding network technology represented by ARPANET was on the verge of being superseded and outcompeted by the proliferation of relatively inexpensive minicomputers. Thus, ‘Had the ARPANET’s only value been as a tool for resource sharing, the network might be remembered today as a minor failure rather than a spectacular success’ (ibid, pp. 104-106). As it happened, email quickly ‘eclipsed all other network applications in volume of traffic’ (Abbate, 1999, p. 107). In fact, according to Hafner and Lyon email amounted to 75 percent of the traffic on the net as early as 1973 (2003, pp. 189, 194). One reason why the ARPANET succeeded as an experiment was that ‘since members used the network protocols in their own work, they had the incentive and the experience to create and improve new services’ (Abbate, 1999, p. 69). That is, as in the case of local email on time-sharing operating systems, network email came as an afterthought, devised by computer technicians for their own use, as a means for coordinating their cooperative effort of building, operating, and maintaining a large-scale construction, in this case the incipient Internet. Email was thrown together like the scaffolding for a new building, only to be- Chapter 11: Formation and fragmentation 361 come a main feature, relegating the resulting building itself, which had been the original and official objective, to something close to a support structure. This pattern — technicians building tools for use in their own laboratories — was to be repeated again and again, as evidenced by, for example, CSNET developed by Larry Landweber and others 1979-81 to provide under-privileged computer scientists access to ARPANET as well as mail, directory services, news, announcements, and discussion on matters concerning computer science; USENET which was developed in 1979 by Jim Ellis and Tom Truscott as a news exchange network for Unix users without access to ARPANET (cf. Quarterman, 1990, pp. 235-251; Hauben and Hauben, 1997); the World Wide Web developed in 1989 at CERN by Tim Berners-Lee and Robert Caillau (Berners-Lee, 1990; Gillies and Cailliau, 2000); and the ARCHIE and GOPHER network file search protocols designed by Alan Emtage in 1989 and Mark P. McCahill in 1991, respectively (Gillies and Cailliau, 2000). 3. The formation of CSCW In important respects, digital computing technologies developed in response to challenges faced by cooperative work. Firstly, the digital stored-program computer was developed as a technology for automating the large-scale calculation work that since the time of the French Revolution had been performed cooperatively, in an advanced form of division of labor (‘human computers’) but by the middle of the 20th century was becoming far too complex to be carried out by the received methods. Secondly, the basic concepts and techniques of interactive computing were developed as a technology of facilitating large-scale cooperative work (air defense, air traffic control, airline reservation) that had become too complex to be done by conventional means, manual or mechanical. Thirdly, the basic concepts and techniques of distributed messaging (network email, etc.) were developed as a technology for facilitating and mediating the coordinative activities required by the cooperative effort of building and managing a large-scale infrastructure, namely, the ARPANET, and were quickly complemented by a host of other technologies such as mailing lists, news groups, file sharing, etc. Furthermore, of course, with the arrival of affordable personal computer (such as the Apple II in 1977, the IBM PC in 1981, and the Macintosh in 1984) and the establishment of the basic Internet protocols (TCP/IP in 1983), interactive computing was reaching a state where it was becoming realistic for ordinary workers, not merely to have access the strictly constrained repertoire of functionality available in real-time transaction processing systems, but to have unrestricted local command of a computer with a palette of applications. It was thereby becoming realistic for computer technologies to become an integral part of cooperative work practices in entirely new ways. 362 Cooperative Work and Coordinative Practices 3.1. Proto-CSCW: ‘Computer-mediated communications’ The new possibilities opened for several partly overlapping research areas with different conceptual frameworks. First of all, of course, email and other technologies of computer-based message handling caught the attention of researchers and technology pundits as a fascinating new topic for studies of communication in organizations, as a new management technology along with the telegraph, teleprinters, fax, microfilm, etc., and engendered a small corpus of evaluation studies of the presumptive effects and impact of electronic mail (Uhlig, 1977; Kiesler, et al., 1984; Eveland and Bikson, 1987) as well as studies that can better be categorized as market analyses (Panko, 1977, 1984). More importantly, a small but persistent research area was formed that focused on developing a computing technology generally known as ‘computer conferencing’ (for an overview of this work, cf. Kerr and Hiltz, 1982). In fact, computer conferencing research developed simultaneously with network email, in the course of the 1970s. But in contrast to email, in a ‘conference’ communications were not restricted to point-to-point message exchanges but were as a default ‘public’ exchanges among attendees at the online ‘conference’. That was the idea, anyway. ‘Computer conferencing’ was originally merely a variant of the online transaction processing paradigm, with dispersed participants logging-in to a central host computer, with ‘messages’ treated like transactions, and with exchanges similarly constrained by a pre-established structure. In EMISARI, for instance, each ‘message’ was confined to 10 lines of text (Kerr and Hiltz, 1982) while control was centralized and fixed. ‘Computer conferencing’ was later, as distributed network architectures became ubiquitous, implemented on a message-passing model; but the notion of centralized control of the communication structure has remained fundamental. In fact, the more ambitious experiments in this line of research, such as EMISARI and EIES by Turoff et al. (Turoff, 1971, 1972, 1973; Hiltz and Turoff, 1978) or FORUM and PLANET by Vallee et al. (Vallee, 1974), explored the rather grand design vision of ‘group communication’ structured according to some presumptively rational model. Later, ‘computer conferencing’ was often advocated as a remedy for the ‘information overload’ which was seen as an inexorable consequence of point-topoint message exchange (Palme, 1984; Hiltz and Turoff, 1985), but, ironically, in the 1980s ‘computer conferencing’ was often used, by users who at the time did not have direct access to the Internet, as a platform for exchanging emails. Anyway, ‘computer conferencing’ soon became categorized together with email as ‘computer-mediated communication’, a label still very much in use (Kerr and Hiltz, 1982; Hiltz and Turoff, 1985; Rice, 1987; Quarterman, 1990; Rice, 1990; Turoff, 1991; Barnes, 2003). However, the research agenda was never succinctly defined but the general drift of the efforts was clear enough. In an early central study of the field, commissioned by the NSF (Hiltz and Kerr, 1981), the authors attempted to ‘collect and synthesize current knowledge about computer- Chapter 11: Formation and fragmentation 363 mediated communication systems’ (ibid., p. viii). However, the ‘synthesis’ focused on issues of ‘acceptance’. The authors note that the idea of computermediated communication ‘seems simple enough at first glance’, but ‘It is the applications and impacts that are startling, and the acceptance of the technology that is problematical’ (Kerr and Hiltz, 1982, p. ix). That is, there was a problem with ‘acceptance’ but the technology was anyhow ‘startling’. Instead of reflecting on underlying design assumptions, the authors argued that, ‘computer-mediated communication […] requires that people accept fairly radical changes in the way they work and even in the way they think, if they are to reap the potential benefits’ (ibid.). In presenting the objectives of their book, Kerr and Hiltz highlighted the following issues in ‘computer-mediated communications’: ‘1. What are the important considerations in designing software or choosing a system from the many available options and capabilities? 2. What factors determine whether such systems are likely to be accepted or rejected? 3. What are the likely impacts of such systems upon the individuals, groups, and organizations which use them? It is not the economic costs and benefits, but the social problems and “payoffs” in the form of enhanced performance and organizational efficiency that should be the main considerations in deciding whether or not to use a computer-mediated communication system.’ (Kerr and Hiltz, 1982, p. x)99 This research agenda has, in a way, become classical: it can be seen as a model of an approach to CSCW research that has been going on stubbornly ever since: which factors determine user ‘acceptance’ and what are the likely impacts of such systems?100 So, while the experiments with conferencing systems sometimes allowed for long-term use and thus evolution of ‘user behavior’ (e.g., Hiltz and Turoff, 1981) and considerable effort was devoted to extensive empirical evaluation, the evaluation studies associated with computer conferencing never were such that underlying principles and concepts of the conferencing idea were questioned: the evaluations invariably focused on ‘acceptance’ and ‘satisfaction’. In any event, this whole line of research on ‘computer-mediated communication’ remained mired in common-sense notions of ‘communication’ with respect to work practices and never began to address the principles and concepts underlying the various communication techniques that had been developed by practitioners and visionaries. The systematic conceptual effort required to transform technique into technology was absent. However, the critical questions were raised, but outside of the small coterie of computer conferencing researchers. The systematic conceptual effort was under- 99 A fourth issue was also listed, but is not really an ‘issue’ but a call for ‘formal evaluation and feedback from users to guide the implementation’ (Kerr and Hiltz, 1982, p. x). 100 Take for example the research agenda of Sproull and Kiesler: ‘How will these technologies influence and change organizations? Does a computer network make work groups more effective? How do people treat one another when their only connection is a computer message? What kinds of procedures best suit long-distance management using a computer network? What problems do these technologies alleviate-and what problems do they create?’ (Sproull and Kiesler, 1991, p. ix). 364 Cooperative Work and Coordinative Practices taken by a separate research program, normally and confusingly also categorized as ‘computer-mediated communication’. The European ‘computer-mediated communications’ community emerged in the wake of the European efforts to develop computer networking (cf. Gillies and Cailliau, 2000). As TCP/IP slowly became available in operating systems and developers began to be able to take it for granted, and as the ‘message handling’ standards stabilized in the first half of the 1980, European researchers, organized under the aegis of the European Commission’s COST-11 program, embarked on what was seen as the logical next step, namely, developing the standards required for putting it all together: email as well as directories, calendars, schedules, and so on. The point of departure for these efforts is well summarized and exemplified by this observation made by Bowers et al. and presented at a European conference on ‘teleinformatics’ (Speth, 1988): ‘Simple electronic mail meets only the most basic messaging requirements of group communication, while relatively sophisticated conferencing and bulletin board systems offering supposedly “all-purpose” facilities reflect their designers’ limited intuitions about what users will wish to do, and such systems can be difficult to adapt to support specific tasks’ (Bowers, et al., 1988, p. 195). That is, as opposed to the previous line of research on computermediated communications, the CMC research undertaken under COST-11 was proactive, predicated on the insight that, for instance, the email technology that had been developed by the enthusiastic ARPANET technicians for their own use was rudimentary. The extant email protocol did not, for example, allow users to configure the protocol for special purposes, nor was the protocol sufficiently expressive for general use in work settings. What was required was a specification language that would allow users to express more than, say, ‘TO, ‘FROM’, and ‘SUBJECT’. Accordingly, ambitious attempts to extend the email protocol by using speech-act theory as a grammar so as to enable users to express the illocutionary point of a message by categorizing it as a ‘REQUEST’ or ‘PROMISE’, cf. THE COORDINATOR by Winograd and Flores (Flores, et al., 1988) and CHAOS by di Cindio, de Michelis, and Simone (De Cindio, et al., 1986). Similarly, it was realized that ‘email’ and ‘conferencing’ should be seen as merely instances of a range of ‘group communication’ facilities that therefore had to be systematically conceptually integrated. Realizing this, research under the COST-11 program aimed at developing such a conceptual foundation. As a representative of one of the research efforts, the AMIGO project, Hugh Smith noted that ‘a number of electronic message systems provide the necessary low-level support for a variety of higher-level structured group communication activities’, such as news distribution, conferencing, information storage and remote retrieval but he then added that ‘there is no direct high-level processing support for these activities within the messaging system’. Specifying a number of critical shortcoming, he emphasized that it was a ‘fundamental requirement that communicated information should be able to be used for more than one function’ but that this requirement was not met, and he went on to note that neither was the requirement that many kinds of services were needed to ‘support structured communication Chapter 11: Formation and fragmentation 365 activities in addition to the basic information transfer capability provided’. He further noted that there were ‘no standardized services to support group communication’, and that most existing systems were ‘inflexible and not easy for the endusers to adapt to their specific purposes’ (Hugh T. Smith, 1988, pp. 90 f.). He called attention to the critical issue that conferencing systems such as those mentioned above ‘often have a centralized resource architecture’ and added that these architectures therefore do not scale up: ‘In the future the sheer size and political/geographical separation of networked communities will demand that the resources and the management of group communication activities be distributed’ (ibid., p. 90). The COSMOS project, also under COST-11, was moving along parallel lines, focusing on developing ‘a high-level, user-oriented language by means of which users can alter the structure of their communication environment’ (Bowers, et al., 1988, p. 195; cf. also Bowers and Churcher, 1988, 125). However, the COST-11 researchers’ critical studies of existing ‘computermediated communications’ techniques and the rigorous attempts to reconstruct ‘computer-mediated communications’ on a systematic conceptual foundation led to unanticipated conclusions: instead of concluding in a systematically reconstructed technology for ‘group communications’, this line of research ended in the realization that the shortcomings they had identified were even deeper. Because of their rigorous approach they realized that problem was rooted in treating ‘communication’ as a separate kind of activity that, presumably, generally is (or can be) carried out divorced from work practices. This realization led to a program shift in CSCW as the focus was moved from the concept of ‘communication’ to the concept of ‘cooperative work practices’. 3.2. The crisis of the message-handling paradigm Although the experiments with ‘computer conferencing’ at the time were reported as very promising and successful, this particular research program ran out of steam. This had to do with these underlying conceptual limitations. ‘Computer conferencing’ research shared with the standard message exchange paradigm the presumption that human communication generally is or can be treated as a distinct activity. True, workers do interrupt their primary work to have conversations and exchange notes, letters, memos about their work (and about other matters). They also, occasionally, put their work aside to go to meetings. For some workers, e.g., managers, the major part of their work day may be spent in conversations and meetings. But apart from managerial work and in the greater scheme of things, conversations and meetings are exceptions, interruptions, ‘a necessary evil’ perhaps, or simply considered ‘a waste of time’. And even when workers engage in conversations and meetings, such discourses are generally related to the state of affairs in their work, to the flow of work, the schedule, the production facilities, and the archives, and in their deliberations workers will discuss schedules, plans, schemes, and so on; they will collate, arrange, distribute, present, hand out, walk up to, gather around, point to, gesture at, inspect, amend, etc. all sorts of artifacts. 366 Cooperative Work and Coordinative Practices By the mid-1980s this insight began to mature and be voiced (cf., e.g., Bannon, 1986, p. 443). The ‘computer-mediated communications’ research program had arrived at a critical junction. The European ‘computer-mediated communications’ researchers soon realized that the ‘message-handling’ model underlying ‘computer-mediated communications’ was quite limited (Pankoke-Babatz, 1989a). In work practices, communication is normally not a separate activity; it is typically an integrated aspect of doing the work. It was therefore considered necessary to be able to incorporate communication functionality in the various domain-specific applications. On the other hand, the European ‘computer-mediated communications’ researchers rejected the ‘computer conferencing’ paradigm as a way to provide structure to the exchange of messaging. Guided by ‘a strong commitment to the actual situation in working life’ (Pankoke-Babatz, 1989c, p. 20), they rejected the idea underlying the ‘computer conferencing’ paradigm of providing ‘a new model’ of communication. Instead, they aimed at providing a model that ‘might be used in the design and implementation’ of local and temporary ‘patterns’ of interaction. That is, instead of deciding on a particular preconceived conception of communication functionalities and applications, they ‘chose […] to look at activities and the regulations required by a group of people to co-operatively execute a particular activity. The model we want to develop should therefore allow specification of such regulations’ (Pankoke-Babatz, 1989c, p. 20). That is, the aim was to build what one could call an abstract model or a notation that would make it possible ‘to model the activities, businesses, tasks, actions or work-flow[s], which are performed by a group of co-operating people’, so as to, in turn, ‘facilitate the required co-ordination and possibly to automate co-ordination, thus reducing the co-ordination effort required of the participants in an activity’ (Pankoke-Babatz, 1989b, p. 14). The European ‘computer-mediated communications’ researchers knew very well that the development of such computational models and architectures would have to be grounded in ‘fundamental understanding of Group Communication processes’ (p. 14), which in turn, because of the complexity and variability of working practices, would need contributions from ‘sociology, anthropology, economics and political science’ (p. 21). Their ‘strong commitment to the actual situation in working life’ was amply demonstrated in the pre-dominance of the practice-oriented program in the European CSCW research community that began to coalesce as these research activities ended in 1988. This critique of the underpinnings of ‘computer-mediated communications’ was also expressed — clearly and succinctly — by Irene Greif in her ‘Overview’ of CSCW in her influential CSCW: A Book of Readings (1988b). Having noted the rapid development of ‘computer-mediated communications’ from electronic mail to computer conferencing she then observed: ‘Computer conferencing has since been expanded to support a wide range of “many-to-many communication” patterns. However, when computer conferencing is applied to some task, the model breaks down. The unstructured body of messages is suitable for the free-flowing text of natural language, but does not let us set the computer to work on our problems. Designers who Chapter 11: Formation and fragmentation 367 draw pictures, software developers who jointly write code, financial analysts who collaborate on a budget — they all need coordination capabilities as an integral part of their work tools. That means coordination support within the CAD engineer’s graphics package, within the programmer’s source-code editor, within the budget writer’s spreadsheet program. It means support for managing versions of objects, be they pictures, programs, or spreadsheets. It means ways to distribute parts of the object for work by contributing group members, ways to track the status of those distributed parts, ways to pull completed objects back together again. The limit of electronic mail and computer conferencing is that they have such features for managing messages only. CSCW widens the technology’s scope of application to all the objects we deal with.’ (Greif, 1988b, pp. 7 f.) Greif’s judgment that ‘the model breaks down’ completely matched the diagnosis that had been made in the European ‘computer-mediated communications’ research community. It is also significant that Greif had reached strikingly similar conclusions with respect to the new research program: ‘Methodologies for testing individual user interfaces don’t apply as well to group support systems. As a result, CSCW is looking more to anthropology to find methodologies for studying groups at work in their natural settings’ (Greif, 1988b, p. 10). In short, it was becoming clear that the ‘computer-mediated communications’ research program was deeply flawed in its underlying ‘message handling’ outlook, in its focus on communication in the abstract, divorced from the work practices of which it normally is an integral part, but also severely limited in the way it conceived of the role of empirical studies in technological development. It was becoming clear, at least to some, that in-depth studies of cooperative work practices in ‘natural settings’ was a prerequisite. The technologies of computer-mediated communication had not failed in any direct sense, nor had their potentials been exhausted. What had happened was what the historian of technology Edward Constant, adopting Kuhn’s concept of ‘anomaly’, has termed ‘presumptive anomaly’. It occurs, not when systems based on the technology fail ‘in any absolute or objective sense’ but when it is assumed or known that the technology in question is seriously limited and that it will fail or be inadequate under certain conditions: ‘No functional failure exists; an anomaly is presumed to exist; hence presumptive anomaly’ (Constant, 1980, p. 15). Take, for example, the turbojet revolution as described by Constant. This technological development did not come about because conventional propeller aircraft technology had ‘failed or faltered by any means or measure: it still held out a great deal of development still to be done; it still promised and in the event delivered greatly increased performance’. But the insights of aerodynamics indicated that the conventional technology would fail when aircraft approached the speed of sound and probably would become inefficient (in terms of speed, fuel efficiency) relative to alternative technologies such as turbojet propulsion (Constant, 1980, pp. 15 f. et passim). The message-handling technologies were seen as having landed in a similar presumptive crisis: on the communication-mediation paradigm, predicated on technologies of message handling, it would not be possible to address the coordination challenges of ordinary cooperative work in a way that integrated communication and coordination with everyday work practices and techniques. It was clear 368 Cooperative Work and Coordinative Practices that message-handling technologies had critical limitations. That messagehandling technologies were found critically limited with respect to work practices does not mean, of course, that they were found useless for cooperative work settings or for other settings. Nor does it mean that message-handling technologies could not be further developed, refined, etc. It just means that they were and remain of marginal relevance to key coordinative issues in cooperative work settings.101 3.3. Automation of articulation work: ‘Office automation’ At the same time as it was becoming clear to many ‘computer-mediated communications’ researchers, especially in Europe, that the ‘message handling’ paradigm was at odds with typical everyday cooperative work practices and that the paradigm thus had to be overcome, researchers in the ‘office automation’ movement were arriving at similar conclusions, although their point of departure was of course entirely different. The ‘office automation’ movement had begun in high spirits in the 1970s, stimulated by different but intersecting technical developments. As with ‘computer-mediated communications’, the baseline was the advent of computer networks. But the approach was radically different. Instead of conceiving of computer networks as a ‘medium’, that is, as a facility that regulates human interaction in negligible ways, the ‘office automation’ program deliberately aimed at regulating interaction in significant ways. The seminal idea was that various new techniques for constructing executable models that had been invented made it worthwhile to explore whether and to which extent such representations might be exploited as a means of modelling and regulating ‘office procedures’ and other kinds of workflows: on one hand, the algebraic techniques for building computational models of distributed systems developed by Petri and others since the early 1960s (cf., e.g., Zisman, 1977; Clarence A. Ellis, 1979) and, on the other hand, the equally sophisticated techniques for constructing complex adaptive models developed under the Artificial Intelligence label (cf., e.g., Hewitt, 1977; Fikes and Henderson, 1980; Barber and Hewitt, 1982). These hopes were soon defeated, however. Experimental applications such as DOMINO turned out to be felt like ‘straitjackets’ in actual use (Kreifelts, 1984; Victor and Sommer, 1989; Kreifelts, et al., 1991a; Kreifelts, et al., 1993). Comparable lessons were learned from the CHAOS experiment (De Cindio, et al., 1988; Bignoli and Simone, 1989; Simone, 1993; Simone and Divitini, 1999). That is, ‘office work’ was not at all as easily captured and modelled as had been presumed. Handling contingencies and dealing with inconsistencies turned out to be an essential aspect of cooperative work practices. The ‘office automation’ program had landed in a crisis of its own. 101 However, this also meant that CSCW as an institutionalized research field would be an arena in which different research programs that do not address the same problem and are characterized by distinctly different research modalities would co-exist. Chapter 11: Formation and fragmentation 369 At this point a new approach to technological research was devised: a few sociologists became involved in the effort to understand the status of ‘office procedures’ and cooperative work in general, on one hand Lucy Suchman and Eleanor Wynn (Wynn, 1979; Suchman, 1982, 1983; Suchman and Wynn, 1984) and on the other Eli Gerson and Susan Leigh Star (Gerson and Star, 1986). That this coupling of sociological and technological research would first occur in the ‘office automation’ movement was hardly accidental. Email and most other communication technologies were devised by computer technicians for their own use. That is, they were developed in a distributed and incremental fashion to solve local problems in practices that were well-known to the designers; and as they were found to be of general utility they were then — post festum — subjected to standardization and design. Their development did not require workplace studies of any kind. On the other hand, computer-conferencing systems were developed in a proactive manner; they were strictly speaking designed. But their design was based on normative models of what was claimed to be rational decision making, not on what was taken to be a well-grounded understanding of an actual practice. By contrast, however unrealistic the experimental designs of the ‘office automation’ movement turned out to be, nobody were under the illusion that one workflow model would fit all, and each workflow model was presumed to be empirically valid. That is, building technical systems that regulate actions and interactions in the strong sense envisioned by the ‘office automation’ movement was unproblematically thought to require some kind of analysis and modelling of existing procedures. When the models ultimately turned out not to work as anticipated, the natural next step was to look more carefully at the reality of ‘office work’. This is anyway what happened. And it was also realized, eventually, that the problem was not just with this or that particular model or modelling technique. It was realized that the problem was conceptual. Those early studies of ‘office work’ indicated that received concepts of cooperative work as mere ‘execution’ of preconceived ‘procedures’ were inherently problematic. This point was driven home, emphatically, both by Gerson and Star and by Suchman in her contemporaneous critique of the concept of ‘plans’ in cognitive science (Suchman, 1987). This insight was a fatal blow to the conceptual basis of the ‘office automation’ movement. 3.4. The CSCW research program The work of Suchman, Wynn, Gerson, and Star had significance beyond these, as it were, immediate implications. It showed, by way of example, that in-depth studies of actual working practices could have strong impact on conceptual issues in the development of computing technologies. It demonstrated the adequacy of, and indeed need for, empirical studies of work practices on the model of Réaumur, Perronet, and Beckmann — even in this area of, putatively, ‘applied’ science. This, in my view, was the defining moment of CSCW. The early contributions by Wynn, Suchman, Gerson, and Star provided the ‘exemplars’, in a Kuhnian 370 Cooperative Work and Coordinative Practices sense, for defining a new research program in which in-depth studies of cooperative work ‘in the wild’ were considered a prerequisite for developing computer technologies for human interaction. However, we should remember that new research paradigms are not necessarily heralded as such when they arrive on the scene. In fact, as pointed out by Kuhn, ‘we must recognize how very limited in both scope and precision a paradigm can be at the time of its first appearance’. Thus the ‘success of a paradigm […] is at the start largely a promise of success discoverable in selected and still incomplete examples.’ (Kuhn, 1962, pp. 23 f.). This observation certainly applies to the emergence of the practice-oriented research program of CSCW. The exemplary role of these studies were not only a function of the findings or of the role of field work in producing them. In both cases the research was integral to settings in which computer scientists and sociologists were addressing the same set of problems. The work of Suchman and Wynn was, of course, an important part of the research at Xerox PARC where Suchman would later head a highly influential interdisciplinary group of researchers. It is less well known but important to note that the work of Gerson and Star anticipated much of was later to unfold in CSCW in that their research was part of a collaborative research network involving both sociologists and computer scientists. The network, which also included Carl Hewitt, Anselm Strauss, Rob Kling, Adele Clarke, Joan Fujimura, Walt Scacchi, and Les Gasser, brought together sociologists with a track record in workplace studies of health care and biological research work as well as computer scientists engaged in developing what would later be known as distributed AI and agent-based architectures. So, when Liam Bannon and I wrote our programmatic article for the first European CSCW conference in 1989, this was the kind of work we had in mind: ‘CSCW should be conceived as an endeavor to understand the nature and characteristics of cooperative work with the objective of designing adequate computer-based technologies. That is, CSCW is a research area addressing questions like the following: What are the specific characteristics of cooperative work as opposed to work performed by individuals in seclusion? What are the reasons for the emergence of cooperative work patterns? How can computer-based technology be applied to enhance cooperative work relations? How can computers be applied to alleviate the logistic problems of cooperative work? How should designers approach the complex and delicate problems of designing systems that will shape social relationships? And so forth. The focus is to understand, so as to better support, cooperative work.’ (Bannon and Schmidt, 1989, p. 360). In sum, two intellectual movements merged in the formation of CSCW. On one hand, ‘computer-mediated communication’ as a technologically oriented research program had arrived at a stage where it was beginning to dawn on many participants that the program was barking up the wrong tree. It had been focusing on aspects of interaction (‘communication’) that were conceived of as divorced from work practices but which normally are an integral part of doing the work and deeply enmeshed in the materiality of the setting and its organizational conventions and procedures. To move beyond that impasse, it was found necessary to develop an understanding of actual cooperative work practices. On the other hand, the ‘office automation’ program had landed in a situation where it had become Chapter 11: Formation and fragmentation 371 clear that formal organizational constructs such as prescribed procedures are not mere algorithmic subroutines but part and parcel of professional work practices. It was, again, found necessary to develop an understanding of actual cooperative work practices. Here the history of CSCW proper begins. A note of clarification. When I point to the early work of Suchman, Wynn, Gerson, and Star as ‘exemplars’ of practice-oriented contributions to technological research, this of course does not mean that the formation CSCW was not part of a wider intellectual movement than circumscribed by Ethnomethodology and Symbolic Interactionism. To the contrary. It was, and is, a distinct research effort within a much broader movement that, in different ways, strives to understand computing in its social context. Thus the Participatory Design movement, which originated in Scandinavia (e.g., Bjerknes, et al., 1987; Ehn, 1988), brought together computer scientists and others striving to understand the development and use of computing systems as embodied social practices (building on the work of Marx, Heidegger, Wittgenstein). Likewise, subversive elements within Artificial Intelligence such as Terry Winograd quite early had serious doubts as to the conceptual foundations of AI and defected, drawing on as diverse philosophical traditions as those of Heidegger and Austin (1986; 1986). In fact, AI was at the then under strong and effective external criticism by philosophers from the same traditions by, e.g., Dreyfus (1979) and Searle (1987, 1989) and from a Wittgensteinian tradition (e.g., Shanker, 1987b, c). At about the same time, a related movement away from cognitive science towards an ‘ecological’ and ‘naturalistic’ conception of computing based on Gibsonian psychology was unfolding in Human Factors engineering (e.g., Vicente and Rasmussen, 1992; Flach, et al., 1995). When I nonetheless point to these early ‘exemplars’ it is because they, in different ways and from different intellectual traditions, demonstrated that in-depth studies of work practices could contribute to the development of the conceptual foundations of computing technology. The fecundity of CSCW’s practice-oriented program became evident immediately, even as the program was being tentatively articulated. The first report on the Lancaster group’s study of air traffic control was presented to the incipient CSCW community in 1989 (Harper, et al., 1989a) and was quickly followed by the equally emblematic study of the London Underground control room (Heath and Luff, 1991a). Nor did it take long for it to become clear that these new insights would have radical implications for not only the development of certain classes of applications but for underlying computer technologies. This was, for example, made explicit with respect to the research area of distributed systems by Rodden and Blair in their classic paper from 1991. Referring to the ‘the rich patterns of cooperation found in CSCW’ as depicted in the early harvest of ethnographic studies, the authors emphasized that coordinative practices are specific to work domain and setting and that cooperating ensembles work ‘in dynamic and unexpected ways and are themselves dynamic’. Having examined distributed systems architectures in the light of this insight, Rodden and Blair concluded that ‘existing ap- 372 Cooperative Work and Coordinative Practices proaches to control in distributed systems are inadequate’ (Rodden and Blair, 1991, p. 49) and that the implications for technological research are fundamental: ‘For example, consider the problem of shared access to resources. In most distributed systems this is dealt with by masking out the existence of other users. Hence sharing is transparent with each user unaware of the activity of others. This clearly contradicts the needs of CSCW. […] The problem with this approach is that presumed control decisions are embedded into the system and hence cannot be avoided or tailored for specific classes of application. This is the root of the problem in supporting CSCW. Because of the dynamic requirements of CSCW applications, it is very unlikely that such prescribed solutions will be suitable.’ (Rodden and Blair, 1991, p. 59). Rodden and Blair concluded that ‘CSCW demands a fresh approach to control which is specifically tailored for cooperative working’ (1991, p. 60). This was a crucial programmatic proposition. The key problem for CSCW is not ‘communication’ or ‘resource sharing’ but cooperating actors’ control of their interaction and, by implication, of the computational regulation of their interaction. The ‘root of the problem in supporting CSCW’, they pointed out, is that coordinative protocols cannot be prescribed once and for all and that ‘control decisions’ must, ultimately, be in the hands of practitioners. This problem is fundamentally different from the issue of user control of system behavior in HCI, in that control in cooperative work settings is, in principle, distributed. In sum, then, informed by the findings of the initial ethnographic studies of cooperative work in real-work settings, Rodden and his colleagues, quite succinctly, articulated a line of research that radically challenges fundamental assumptions in computer science as expressed in the architecture of operating systems, network services (client-server protocols), database systems, etc. The ‘root of the problem in supporting CSCW’ has since then been spelled out and explored from different perspectives: ‘event propagation mechanisms’ for ‘awareness’ support, ‘coordinative artifacts and protocols’, and so on. With the initiation of this line of research, the CSCW research program had so to speak been fully specified. The early studies by Wynn, Suchman, Gerson, and Star had demonstrated how sociological inquiries could address conceptual issues in technological research. The potential of the practice-oriented program, as exemplified by these early paradigms, had been decisively demonstrated in the studies of control room work and other studies that then made Rodden and his colleague subject distributed systems theory to critical scrutiny. Furthermore, the work of Rodden and his colleagues also had significance beyond its immediate implications for computer science. With their re-conceptualization of fundamental issues in distributed computing, CSCW’s practice-oriented research program had been complemented by an exemplar of the correlative technological research. The reciprocality of the contributions of sociology and computer science respectively had been exemplified. In sum, what had been articulated was a technological research program devoted to the development of computing technologies — but not just that: it was a quite special technological research program, namely, a technological research program that depended critically on conceptual contributions from sociology and Chapter 11: Formation and fragmentation 373 anthropology. That is, the required empirical work was proactive in its orientation: intended to contribute to the conceptual foundation of the technology. In this regard, CSCW is quite extraordinary: technological development predicated on ethnographic studies This research program thus differed distinctly from the kinds of sociological and socio-psychological studies of technology that had been conducted previously, in that the emphasis was on development of technology, not on evaluation of ‘user satisfaction’ with particular systems, and even less so on prognostications about the organizational or behavioral ‘effect’ and ‘impact’ of certain technologies. 3.5. Technology and ethnography: An ‘odd mix’? CSCW seems like a strange bird. To some it looks like ‘requirements engineering’, ‘systems development’, or participatory design. Others, however, simply take it to be a special area within HCI. To others again it seems like a version of technology assessment or quality evaluation, or even an extension of media studies. What causes the confusion may be the seemingly paradoxical combination of technology development grounded in, of all things, ethnographic studies of work practices: an ‘odd mix’ indeed! (Finholt and Teasley, 1998). The sense of ‘oddity’ is dispelled when CSCW is looked at in the light of history of technology. As pointed out above, technological research and development is immensely variegated. Techniques are generally developed in an ‘empirical’ process of incremental improvement of existing techniques, typically in reaction to known or anticipated bottlenecks, problems of reliability, performance limitations, production and maintenance costs, etc. The same, of course, applies to technologies, only that, with technologies, the development involves systematic analysis and rationalization of existing techniques and technologies. Still, technologies have systemic character; they form complexes of component and enabling technologies. Fortuitous arrival of new component or enabling technologies may suddenly propel a given technology forward. The microprocessor is a case in point. It was developed for desktop calculators but was then laterally shifted and appropriated as an essential component technology for personal computers. Now, some technologies are developed in a proactive way, in an open but directed search for solutions to deal with a (known, anticipated, or imagined) societal problem or to create new possibilities. Obvious examples are railway technology, telegraphy, radio communication, turbojet, radar, space flight, satellite navigation, nuclear fusion energy — as well as digital electronic computing, real-time transaction processing and digital computer networking. In its orientation, CSCW belongs to this family of technology development efforts. That is, in its orientation, the CSCW research program runs counter to the research program that dominates HCI. HCI represents a quite normal modus of technological research in computing technologies and in general. HCI research arrived on the scene after interactive computing, and it was from the beginning 374 Cooperative Work and Coordinative Practices devoted to the conceptual rationalization and refinement of technologies for which the basic principles had already been developed, albeit not systematically articulated and conceptualized. But for CSCW’s practice-oriented research program, the requisite technology — the basic concepts and principles — hardly exist yet but have to be conceived and developed in a proactive technological research effort informed by ethnographic and other forms of workplace studies. That is, the two research fields, although closely related with respect to many component technologies but also, to some extent, with respect to the disciplines and methods involved, represent radically different modalities of research. This difference in research modus has been and remains a source of much confusion. To make the confusion worse, and as shown above, email and other forms of ‘computer-mediated communication’ developed in a ‘spontaneous’ or ‘empirical’ (unintentional, distributed) process of innovation and dissemination made possible by the fact that technicians ‘used the network protocols in their own work’ and therefore had ‘the incentive and the experience to create and improve new services’ (Abbate, 1999, p. 69). Thus, just like HCI, CSCW was anticipated by decades of incremental innovation of computer-mediated communication techniques such as email, instant messaging, shared data sets, and so on, and to continue in this modus would therefore easily appear the natural and unproblematic way of proceeding. However, CSCW’s practice-oriented research program reversed CSCW’s research modus in the course of the early years of formation of CSCW. The practice-oriented CSCW research program is oriented in the opposite direction vis-à-vis practical technical innovation. There is yet another source of confusion, however. The design of the very first computer applications for commercial purposes (payroll systems, etc.) were based on studies of actual practices. As early as 1953, the requirements analysis for one of the first business applications, the design of a program for the ordering of goods for Lyons Teashops in the UK, involved genuine field work (Ferry, 2003, pp. 121-129). What was new in CSCW was not the idea of doing requirements analysis as an integrated part of the process of building a particular system for a particular setting, incorporating an array of more or less well-known technologies in a way that serves the identified needs. What was new in CSCW was the idea of doing workplace studies for the purpose of developing new technologies. In-depth studies of coordinative practices in the age-old tradition exemplified by the ‘dry, thorny, and not at all dazzling’ studies conducted two or three centuries ago by French academicians such as Réaumur and Perronet or by German scholars such as Beckmann and Poppe are critically important for CSCW. In that respect, there is nothing ‘odd’ in technological research informed by studies of work practice. However, the specific function of these studies in CSCW is not the improvement of the investigated practices in any immediate or direct sense102 but 102 Studies of coordinative practices with a view to their systematic rationalization would fall under the rubrics of operations research, logistics, etc. (e.g., Morse and Kimball, 1951). Some of the work of ‘scientific management’ practitioners also belong here. On the other hand, ‘Business Process Reengineering’ is too Bolshevistic in its general stance towards actual work practices to be considered in this context. Chapter 11: Formation and fragmentation 375 the development of novel coordination technologies that may then be appropriated for the investigated practices (as well as others) in order to improve and transform their practices. In other words, in contrast to the classic technological studies of handicraft-based work practices, the practice-oriented research program of CSCW is not reactive (post hoc systematization and rationalization) but essentially proactive. In CSCW, the path from ethnographic studies to the conception of technical solutions to systems design is far more convoluted than in the classic technology studies and far more convoluted than suggested by the notion of ‘implications for design’ (Dourish, 2006b; Schmidt, et al., 2007). 4. Accomplishments and shortcomings It is often intimated, if not indeed insinuated, that CSCW has not accomplished anything substantial. Or it is bluntly alleged that CSCW has not yet resulted in technologies in actual use. The situation is more nuanced than that. Take Lotus Notes, for example. Developed in the late 1980s and early 1990s by a team at Lotus headed by Irene Greif, and originally released in 1989, it not only had a certified CSCW pedigree but also incorporated important early CSCW insights concerning the importance of integrating the provision of message-handling functionalities such as email and computer conferencing with a shared document database system over the Internet (Greif and Sarin, 1986; Kawell, et al., 1988). More importantly, its example spawned a large and rather messy market for what is sometimes called ‘collaborative working environments’ or ‘integrated groupware platforms’ (e.g., Lotus Notes/DOMINO Microsoft SharePoint, Novell SiteScape, etc.). By 2000, about 70 different commercial platforms of this kind had been registered, and at about the same time more than 19,000 software houses world-wide were offering products for the Lotus Notes/DOMINO platform alone (Wilczek and Krcmar, 2001). It is estimated that the global market for these systems by 2005 was in the magnitude of $2 billion (Prinz, 2006). The number of users of such systems is estimated to be well above 200 million, a number that does not include Google’s Mail, Wave. Calendar, etc. So, if market size and actual use are the criteria, then something must be said to have been accomplished. But in fact, the vast majority of what is being marketed as ‘collaborative environments’ represents the state of CSCW by the end of the 1980s, that is, prior to the systematic coupling of workplace studies to the development of technologies could have an effect. Even more so, the design principles underlying these systems — one could call it the ‘groupware’ paradigm — derives from the prehistory of CSCW, not from ethnographic work in CSCW. ‘Groupware’ is basically defined as a package of computer-mediated communication functionalities such as email and instant messaging combined with a shared repositories. These basic functionalities are now generally supplemented by calendar and address book functions. These functionalities are to some extent integrated, typically by offering elementary interoperability of calendar and mes- 376 Cooperative Work and Coordinative Practices saging in the form of alarms, but also shared address lists, and so on. ‘Groupware’ is a category of products that have be shaped with a view to a mass market that not only includes team communications in work settings but also any setting where people may form teams or groups. It is as close to a generic product as it can get. ‘Groupware’ packages are expressions of a design strategy that offers integration of messaging and file repositories within the framework of existing operating systems architectures, that is, as yet another application next to other applications. From the point of view of commercial software design, this may be rational, for one thereby dodges the challenge of developing computational coordinative protocols that can be invoked by any application. The point is that in the design of these products the thicket of cooperative work practices has been carefully avoided. The technical solutions to the problems CSCW is addressing will have to be incorporated in all kinds of software machines, from operating systems and network protocols to domain-specific applications (desktop publishing, computer-aided design, production planning and control, electronic patient records, and so forth). CSCW and Groupware are not coextensive fields. The conceptual horizon of groupware ends where that of the CSCW research program begins. 4.1. Ethnography and technology: A ‘big discrepancy’? When allegations about CSCW’s lack of fecundity are being made, what is implicitly expected is that what is unique to CSCW should have led to novel technologies, namely, the systematic coupling of ethnographic and other forms of indepth workplace studies to the development of technologies for cooperative work settings. And the impression is that this coupling is not working properly. A widely cited source of the impression that there is such a problem, a ‘gap’, in the transmission of insights from workplace studies to ‘design’ is an article by Plowman, Rogers, and Ramage in which they reported on a survey of a large part of the workplace studies that had by then been published in the area of CSCW (1995). In the article, the authors asserted to have found a ‘paucity of papers detailing specific design guidelines’ (p. 313). They carefully and explicitly refrained from concluding that ‘workplace studies do not produce specific design guidelines’, but nevertheless suggested that the observed paucity ‘can be attributed to the lack of reported research which has developed to the stage of a system prototype’ (ibid.). They went on to surmise that what impeded the progression from workplace studies to ‘the stage of system prototype’ could be what they described as ‘a big discrepancy between accounts of sociality generated by field studies and the way information can be of practical use to system developers’ (p. 321). Not surprisingly, this conjecture (and others in the same vein) has fuelled widespread concern and continual discussion about the role of workplace studies in CSCW (Dourish and Button, 1998; Crabtree, et al., 2005; Dourish, 2006b; Randall, et al., 2007a; Crabtree, et al., 2009). Chapter 11: Formation and fragmentation 377 It is undoubtedly true that the sample analyzed by Plowman, Rogers, and Ramage showed only few papers ‘detailing specific design guidelines’. However, their assertion that there is a ‘lack of reported research’ informed by workplace studies ‘which has developed to the stage of a system prototype’ was false. As pointed out in my contribution to the collection on Workplace Studies edited by Luff et al. (2000), the alleged ‘paucity’ does not ‘reflect on the actual impact of workplace studies on the development of CSCW technologies’ (Schmidt, 2000, p. 146). It is simply an artifact of a flawed method: The ‘histories of how workplace studies inform the development of CSCW technologies’ are ‘not always readily visible in papers reporting on findings from workplace studies. The transfer of findings and insights typically happens in the course of discussions within crossdisciplinary research teams and are often only documented in design-oriented papers’ (Schmidt, 2000, p. 148, n. 3). In other words, the presupposition that there is an immediate and direct route from workplace study to ‘implications for design’ to ‘detailed design guidelines’ to the ‘stage of system prototype’ reflects a widespread but simplistic notion of the development of technology. To investigate the development of knowledge in any field, and most certainly the development of technology in an interdisciplinary field such as CSCW, would require that one traces and maps out the complex pattern formed by the more or less tentative articulations, interpretations, categorizations, conceptualizations of observed phenomena; the ways in which concepts and ideas percolate and propagate within the research community, as manifested in more or less sporadic citation patterns or in the often elusive cross-fertilization among researchers in collaborating laboratories and research projects; and the various simultaneous or subsequent generalizations and transformations of these concepts and ideas as they are appropriated by other actors, perhaps pursuing diverging aims and addressing different problems and issues. Thus, in the case of the impact of workplace studies in CSCW on technological development one would have to investigate the (partial and patchy) conceptual framework emerging from the propagation and transformation of concepts derived from workplace studies, such as ‘awareness’, ‘boundary object’, ‘artifact’, ‘work setting’, ‘coordinative protocol’, etc. and how this (partial and patchy) conceptual framework is being interpreted and appropriated in different ways in technological research. There is another source of ambiguity in determining the form and extent of the impact of workplace studies on technical research in CSCW. Workplace studies in CSCW play two major roles. On one hand, workplace studies have served the critical role of demonstrating the ‘socially organized’ character of human activity. This critical role is obvious and has been the cause of justifiable celebration. For instance, in 2002 Jack Carroll published a collection of articles (under the title Human-Computer Interaction in the New Millennium (Carroll, 2002) to take stock of the accomplishments of HCI and discuss the challenges and opportunities facing the field, in which he makes the following observation: ‘At first, the focus [of CSCW research] was on collaborative systems (groupware) and humancomputer interactions with collaborative systems. But the more significant impact of CSCW, one 378 Cooperative Work and Coordinative Practices that goes beyond the conference, is the recognition that all systems are used in a social context. 1n this sense. CSCW has become more a view of HCI than a subcommunity with[in] it. CSCW has served as a conduit for the expansion of the science foundation of HCI to incorporate activity theory, ethnomethodology, and conversation analysis, among others’ (Carroll, 2002, p. xxxiii). This is a distinguished accomplishment by any standard. On the other hand, however, workplace studies also serve the constructive role of being instrumental in providing, inductively, the empirical and conceptual basis for the technology development that CSCW is ultimately all about. This role is much less obvious and not at all well understood. 4.2. The case of ‘awareness engines’ To take but one example: the concept of ‘awareness’. As CSCW formed as a research field, it was quickly established that the coordination and integration of cooperative work activities typically involve highly sophisticated ‘awareness’ practices, i.e., practices of ‘monitoring’ unfolding occurrences in the setting and, conversely, of making local occurrences ‘publicly visible’ for others in the setting (e.g., Harper, et al., 1989a; Heath and Luff, 1991a). These practices, that are generally named ‘mutual awareness’, are crucial because they are practically effortless. Although they involve delicate techniques, they typically require no noticeable effort of skilled practitioners because they are spatio-temporally integrated with the activities of doing the work. Skilled actors are able to coordinate and integrate their interdependent activities without having to interrupt the flow of work. In fact, to be able to do so is typically what is required to be considered competent in a given cooperative work practice. This insight prompted intensive research into different ways of providing computational support for these subtle practices, and over the years different approaches have been developed and explored (for an overview, cf. Schmidt, 2002b). However, what is interesting here is, first of all, that this rich line of research, while inconclusive, has resulted in facilities that are in actual use. The obvious instance of this is the early web-based groupware system named Basic Support of Cooperative Work (BSCW), initially developed at GMD in Germany in the mid 1990s. In contrast to the line of ‘collaborative environments’ in the Lotus Notes tradition, BSCW offers elementary support for ‘mutual awareness’. Users can see traces of actions on documents in the workspace in which they are participants. Each user can, for example, see if and when another user has ‘read’ a document, changed its name and description, moved it, and so on. Not only is BSCW of CSCW provenance; the design of the system was clearly informed by the findings of early ethnographic studies (Mariani and Prinz, 1993; Bentley, et al., 1997a; Bentley, et al., 1997b; Appelt, 1999). BSCW is currently installed on more than 1,000 servers, typically at research institutions, servicing ‘tens of thousands of workgroups and hundreds of thousands of users’ (Orbiteam, 2009). Log-analysis studies of patterns of usage in BSCW show that the ‘awareness service’ is widely used, especially by experi- Chapter 11: Formation and fragmentation 379 enced users (Appelt, 2001). Nevertheless, the ‘awareness service’ technology as represented by BSCW is crude and primitive. Its major limitation is that the massively distributed nature of awareness practices is severely curtailed by the architecture. The shared ‘workspace’ (i.e., a directory) is a distinct set of objects to which a user obtains access by invitation (although ‘rights’ may differ), and events thus cannot propagate beyond the ‘workspace’ as defined by the inheritance rules of the directory structure. In shot, BSCW’s ‘awareness service’ should merely be taken as an inspired first short at how computational artifacts might be designed to support the observed ‘awareness’ practices (for a critical analysis of BSCW, cf. Schümmer and Lukosch, 2007, pp. 501-525). The picture of ‘awareness’ practices that emerged from the initial ethnographic studies, was one of horizontal adaptation in a massively distributed mode of interaction. Cooperative work is essentially characterized by distributed control. Now, work plans and conventions and the concomitant coordinative artifacts of course play a critical role in reducing the effort of handling the complex interdependencies by providing the normative background of taken-for-granted expectations (‘He’s supposed to have finished x by now, and… yes indeed, there it is!’). But they can be seen as ‘local and temporary closures’, to use the apt expression of Gerson and Star (1986), in contrast to which ‘awareness’ practices can very well be understood as ‘the work to make work work’. The importance attributed to the concept of ‘awareness’ in CSCW research largely derives from this insight, and there is a close conceptual affinity between concepts like ‘situated action’, ‘articulation work’, and ‘mutual awareness’. That is, the defining feature of ‘awareness’ practices as the concept emerged from the ethnographic record is that of massive distributedness. In direct continuation of these findings, CSCW research into how to ‘support’ awareness practices has increasingly and predominantly pursued calculi of distributed computation so as to be able to express, facilitate, emulate collaborating workers’ practices of paying heed to occurrences in their work settings. An early example is the ‘spatial’ or ‘aura-focus-nimbus’ model that was developed by Benford and others in the COMIC project (Benford and Fahlén, 1993; Benford, et al., 1994a; Benford, et al., 1994b) and was later generalized by Rodden (1996) and Sandor et al. (1997). Another approach, suggested by Simone and Bandini (1997, 2002), explored the possible utility of the reaction-diffusion model as a way of representing and facilitating the distributed propagation of state changes within a cooperative work setting. What characterizes these otherwise very different approaches is that they all, in the words of Sandor et al., attempt ‘to integrate awareness support at fundamental levels of cooperative system architecture’ (Sandor, et al., 1997, p. 222). As noted above, this approach was explicitly ‘informed’ by the ethnographic record (cf., e.g., Rodden and Blair, 1991; Rodden, et al., 1992). In other words, it was realized at a very early stage in CSCW that, since ‘awareness’ practices are an integral aspect of cooperative work practices, ‘awareness support’ cannot be provided by a special kind of application, on par with, say, email or instant messaging; it 380 Cooperative Work and Coordinative Practices has to be provided as a service available to all applications. It was therefore realized that ‘awareness support’ would have to be provided ‘at fundamental levels of cooperative system architecture’, that is, at the level of the operating system of networked computational artifacts. In short, the early ethnographic studies were almost immediately seen to have radical ‘implications for design’ with respect to operating systems architectures, database technology, and so on.103 The ‘groupware’ model was no longer viable as a model for CSCW. This has further implications. It means that the transition from workplace studies to the development of new technologies for, e.g., ‘awareness support’ would become quite complex. Major conceptual and technical issues had — and have — to be identified and resolved, and at the same time the enormous investment in the installed base of computer platforms and application software poses a major buffer against anything like swift change. This in turn has had the effect that ‘awareness support’ provided by IT systems in actual use has stayed at the level represented by the rather crude facilities of BSCW or the even cruder ‘awareness support’ offered by instant messaging where users are notified (by icons or sounds) of the on-line presence of ‘buddies’. The case of ‘awareness support’ is illuminating in that it also demonstrates internal sources of seeming stagnation. The trajectory of the line of technological research on ‘awareness support’ that is informed by ethnographic studies is strikingly convoluted. What was initially being pursued was a calculus for modelling massively distributed interactions among entities. It goes without saying that facilities for integrating ‘awareness support’ at fundamental levels of operating system architectures will have to be quite generic, but these facilities will at the same time also have to be able to support the domain-specific character of these practices. The ethnographic finding that practices of heeding (to use a less misleading term than ‘awareness’) are domain-specific practices seem to have been put aside, temporarily perhaps, in favor of advancing the ability to model distributed social processes in general. In cooperative work, actors skillfully ‘monitor for’ certain cues, signals, etc. in the setting and, conversely, skillfully ‘display’ certain cues, signals, etc. concerning their local activities as relevant to colleagues. That is, practices of heeding in cooperative work is first of all not a mental state acquired ‘passively’ (as the term ‘awareness’ might suggest) but a practical stance of monitoring (or displaying), and the categories of cues, signals, etc. that actors are monitoring for (or displaying) are specific to the given domain of work and the work setting. Practices of heeding are not generic abilities; they are skills practitioners only master with training and experience, highly elaborate practices that are virtually effortless because they are an integral aspect of the domain-specific work practices. And like any other practice, they are essentially conventional and thus normative (‘What are you doing?! Are you asleep?’ or ‘Don’t worry; I see 103 Identical conclusions have been drawn from other workplace studies investigating ‘coordination mechanisms’ in cooperative work settings (Schmidt, et al., 1993; Schmidt, 1994d). Chapter 11: Formation and fragmentation 381 the problem and can handle it!’). Without facilities that support actors in these skilled techniques, by allowing them to express the categories of cues, signals etc. they monitoring for (or displaying), in short, the protocols of their heeding practices, computational support of ‘mutual awareness’ is unlikely to proceed significantly beyond the rudimentary ‘awareness service’ of BSCW or similar. There may have been sensible intellectual economy in putting these issues aside. It is interesting, however, that recent research has begun to address those very issues. Gross and Prinz have for example presented an ‘Event and Notification Infrastructure’ that aims at providing ‘awareness information in a way that is adequate for the current situation of the user’. Gross and Prinz emphasize that ‘context itself depends on parameters like the current task, the current type of cooperation, the artefacts and tools used, and so forth’ (Gross and Prinz, 2003, p. 296). In fact, recognizing that heeding practices, like any other practice, are essentially normative, this research aims at developing facilities for contextual awareness support that ‘will allow users to establish conventions’ (Gross and Prinz, 2004, p. 300). This brings us back to the problem of the putative ‘gap’ or ‘discrepancy’ in CSCW. The problem is not that there is a ‘lack of reported research [informed by workplace studies] which has developed to the stage of a system prototype’. There is evidently no such lack; in fact, the ethnographic findings have obviously engendered a rich tradition even in the fairly esoteric area of ‘awareness support’. This should not be taken to imply that the relationship between workplace studies and development of technology is unproblematic. Workplace studies are no substitute for serious and meticulous conceptualization. On the contrary, and to paraphrase Garfinkel, workplace studies in CSCW serve ‘as aids to a sluggish imagination’ (Garfinkel, 1967, p. 38). The function of workplace studies is not to produce ‘implications for design’ or anything of the sort but first of all to challenge taken-for-granted assumptions about cooperative work and coordinative practices and thus kindle an otherwise sluggish technical imagination. Thus, bringing findings from ethnographic studies of cooperative work to bear on technological development involves conceptual work — or rather: it essentially consists in conceptual work. As suggested above, there are two aspects to this: a critical one and a constructive one. The conceptual work of bringing findings from ethnographic studies of cooperative work to bear on technological development involves dissolving run-of-themill constructions seeping in from the various disciplines and from the metaphysics of common-sense theorizing (e.g., ‘shared goals’), questioning what is inadvertently taken as prototypical (e.g., ‘face-to-face’ interaction in ‘teams’), sorting out category mistakes and the ubiquitous transgressions of sense (e.g., ‘media richness’), and so on. However, the research on ‘awareness support’ stumbled in this critical effort. The early ethnographic studies did not interpret the findings in terms of ‘awareness’. In fact, the term ‘awareness’ doesn’t appear there. The findings were nonetheless soon interpreted in the light of that concept. Now, since ‘awareness’ 382 Cooperative Work and Coordinative Practices is an ‘attention word’, not a ‘heed word’ (Ryle, 1949; White, 1964), this caused significant confusion. The ethnographic findings were implicitly given a mentalistic interpretation by being subsumed under the notion of ‘awareness’ that seems to have been imported from social psychology and small-group sociology. The findings produced in the early ethnographic studies were thereby, so to speak, contaminated by abstract constructs, obviously modelled on so-called ‘face-toface interaction’, that had more affinity to Goffman’s notion of ‘focused’ and ‘unfocused’ interaction (Goffman, 1961, 1963) than with the highly specific findings articulated by the early ethnographic record. The result was a mentalistic notion of ‘awareness’ from which actual practices and protocols had somehow been excluded. The mentalistic interpretation in turn encouraged design experiments such as ‘media spaces’ in which much effort was spent on trying to recreate, with as much fidelity as possible, the putative paragon of all human interaction: the ‘face-toface’ chat in the office corridor (for critical analyses, cf. Gaver, 1992; Heath and Luff, 1993). As a result, much of the crucial insight gained by the ethnographic findings was easily overlaid by concepts from social psychology such as ‘shared understanding’, ‘shared goals’, etc. in which the very practices through which ‘understanding’ or ‘goals’ become ‘shared’, i.e., unproblematically aligned, taken-forgranted, etc., are glossed over. Consequently, many of the initial explorations into ‘awareness’ services were experiments with — exactly — shared data sets facilitated by joint access to a directory on a server, perhaps augmented by a notification service. Sometimes this was graphically dressed up in a ‘room’ metaphor. However, it was quickly realized that the conception of cooperative work arrangements underlying this design was faulty, predicated as it was on the deeply problematic idea of the ‘group’ or ‘team’ as a well defined and clearly bounded unit. Researchers accordingly began to deconstruct the notion of cooperative work arrangements as defined in terms of physical space (cf., e.g., Fitzpatrick, et al., 1996; Harrison and Dourish, 1996; Dourish, 2006a; Ciolfi, et al., 2008). Bringing findings from ethnographic studies of cooperative work to bear on technological development also requires the constructive task of reconstructing the logic of ‘awareness’ practices, the protocols or conventions that competent actors routinely follow and expect others to follow, and the ways in which such protocols or conventions are established and maintained. This kind of investigation cannot be driven by the way sociology or anthropology frames problems. To be able to ‘inform design’ this kind of investigation must be informed by the way technological research tentatively frames its problems. It is in this context noteworthy that another group of researchers, led by Carla Simone, in pursuing what seems to be the exact same research problem, has resolved that the ethnographic evidence of heeding practices that was produced almost twenty years ago (in a different intellectual milieu and in pursuit of far less specific research questions), now needs to be complemented and enriched. Simone and her colleagues have therefore undertaken new ethnographic studies of awareness practices with an ex- Chapter 11: Formation and fragmentation 383 plicit focus on the ‘conventional’ or normative nature of these practices (Cabitza, et al., 2007; Cabitza and Simone, 2007; Cabitza, et al., 2009). The point of all this, then, is that the transition from ethnographic studies to technological development is an immensely complex effort. On closer inspection, what may look like a ‘gap’ turns out to be a maze of pathways. Some lines of research lead straight to useful and innovative applications. Other lines of research turn out to be futile. Yet other lines of research turn out to raise questions that require additional ethnographic work. And so on. That said, while there is no ‘gap’ or ‘discrepancy’ between workplace studies and technological development, the impression that something is amiss in CSCW is not entirely wrong. I suggest that the perceived sluggishness, the impression that progress is at best intermittent and hesitant, at worst that things are going in circles or backwards — to a large extent is precipitated the an increased fragmentation of CSCW. 4.3. Logics of fragmentation There is an interesting logic to the eventual fragmentation of CSCW. As always with any kind of research, the presence of distinctly different paradigms is of course a recipe for trouble. CSCW is certainly no exception. More specifically, the continued presence of research within CSCW conducted on the ‘groupware’ paradigm would itself make the very concepts of ‘work’ and ‘work practice’ problematic. In a ‘groupware’ perspective the term ‘work’ does not mean work practice, for that concept is alien within a ‘groupware’ paradigm; but rather, the term ‘work’ merely stands proxy for ‘workplace’ or ‘team’, for what is addressed is a priori delimited to communication in abstraction from the specific skills, techniques, procedures, and material settings involved in work. In ‘computer-mediated communications’ or ‘groupware’ systems research, ‘work’ as a research phenomenon has been transubstantiated, in much the same way as ‘work’ is spirited away from much of ‘sociology of work’ (Sharrock and Anderson, 1986, Chapter 6). It was therefore but a small step for ‘groupware’ research to abandon the pretense that this research has anything to do with ‘work’ in any ordinary sense. On the other hand, however, the practice-oriented research program crumbled from within. It has made significant progress but has not succeeded in showing convincing technical solutions to the problem of computational regulation of cooperative work in complex settings. This is not surprising. In contrast to ‘groupware’ research and development, the practice-oriented research program is not about developing a new class of applications on par with other applications but is about developing technologies that will provide all applications with coordinative functionality. That is, CSCW has to break the ‘groupware’ barrier. What is required is technologies to support workers in coordinating their cooperative activities in the domain-specific categories of their work and from within their domainspecific application programs, and the coordinative functionality thus has to be 384 Cooperative Work and Coordinative Practices provided by computational artifacts that — as separate control mechanisms residing in the network operating system — have to interface with the data structures and transformation processes of the application programs. This applies to ‘awareness support’ as well as ‘workflows’ or ‘ontologies’. To put it directly, CSCW research is developing a new kind of technology that, compared to the large installed base of polished ‘groupware’ products, does not have much to show for it — not until all the elements are there and have been put together, that is. But these problems of public perception are the least of it. The real issue is that the CSCW research program was never clearly articulated. Great progress was achieved in involving workplace studies in technological research, but at the same time the very notion of work practice has been vacated, emptied of other content than the notion of mere contingent activity, disconnected from the concepts of rules and regulations. The practice-oriented program of CSCW — the very idea of computational regulation of interaction — has ended in contradiction. On one hand ‘groupware’ is seen as severely limited because it is predicated on communication divorced from work practice, but on the other hand the integration of computational functionality into the coordinative practices is seen as conceptually wrongheaded. This contradiction has, in the end, made the CSCW research program largely paralyzed. Critical workplace studies certainly have flourished and have been justly celebrated, but constructive workplace studies, goal-directed studies of coordinative practices, protocols, logics of combination and recombination, have only progressed slowly and sporadically. So, when frivolous applications of computing technologies became a mass market and seized the imagination of the public, with such frail foundations, the practice-oriented research program was a push-over. 385 Chapter 12 Frail foundations CSCW, as a research area, is in disarray. Not only are there different schools of thought, but the different communities are not investigating the same phenomenon or the same kind of phenomena, nor do they engage in any kind of discourse about findings. In fact, they would not even be able to compare notes. This would be a serious predicament for any research area. For an interdisciplinary area this is fatal. What normally unites a research area or a discipline is the common framework of exemplars, methods, techniques, textbooks, educational programs and institutions, etc. In an interdisciplinary field, all this is missing, for here all or most of this scaffolding belongs to the constitutive disciplines. What unites an interdisciplinary area is, by and large, merely the research problem its members have gathered to investigate in their different ways and the ‘cornerstone’ concepts in which the problem is identified and expressed. But if the very conceptual foundation, the set of cornerstone concepts, is muddled, incoherent, disputed, then none of the other factors that otherwise assist an area in conceptual crisis to get through to the other side are there to stabilize the area. Accordingly, without reasonable clarity about CSCW’s conceptual foundation, its research problem, the research community has no accountable criteria of quality, relevance, priority, directions. Its research program dissolves; surpassed paradigms linger on among the living; arguments erupt about shifting the area’s focus this way or that way or even ‘widening’ it. CSCW is therefore subjected to centrifugal forces that are tearing it to pieces: disciplinary chauvinism, distractions of shifting funding schemes, changing technical fashions, frivolous media interests, etc. Why is CSCW is such disarray? Why has CSCW been unable to effectively supersede the various lines of research focusing on mediation of communication? Why has there, in fact, been a regression to versions of ‘computer-mediated communications’ that focus on evaluation studies of new applications of the wellknown technologies of exchanges of messages and files? There are several reasons for that. One is surely that CSCW researchers have been reluctant to pursue the design of computational artifacts for regulating cooperative work and thus are inclined to conceive of ‘Computer-Supported…’ as something close to ‘Computer-Mediated…’. Their motives may be different: they may, based on experience, 386 Cooperative Work and Coordinative Practices find existing coordination technologies of this sort intolerably rigid; they may be inhibited for ideological reasons; or they may be of the persuasion that regulation by machines of human activity inherently impossible, a conceptual chimera. In fact, the motives may be mixed, but the latter seems to be the one that has carried most weight. Reluctance to pursue the design of computational artifacts for regulating cooperative work may also go a long way to explain the urge to the move the focus of CSCW ‘away from work’, for in leisure and domestic settings there is no compelling need for regulation of interaction by means of sophisticated coordinative artifacts and, hence, neither by means of computational artifacts, and even less so is there a need for actors to modify and construct computational artifacts to regulate interaction. Such practices typically belong to professional work in complex settings. And on the other hand, communication technologies such as email, http, messaging, chat, etc. are not domain-specific; they do not even have a domain bias. They are as generic as pen and paper. That is, when the focus is on these generic communication technologies there is no reason whatsoever for focusing on cooperative work. Hence, I suggest, the fragmentation of CSCW. That is, there is deep confusion in our understanding of work practice, of action and plan, of computation, and hence of the very issue of computational regulation of cooperative work. In short, there is considerable confusion in CSCW concerning the field’s conceptual foundation. The conceptual confusion is double sided: on one hand, the notion of ‘plans’ or ‘procedures’ versus ‘situated action’, and on the other hand, the notion of ‘computation’ inherent in any conception of computational regulation. These notions are all fraught with confusion and mystification. I made a modest attempt to elucidate the concepts of ‘plans’ and ‘situated action’ many years ago (Schmidt, 1994d, 1997), albeit with little apparent success. Suchman, for instance, never dignified the critique with a reply (cf., for example, Suchman, 2007). There may be several reasons for this. Some would, for example, argue that I misunderstood Suchman’s argument and made a fuss about nothing. Alternatively, though, my argumentation may have been too hushed to be heard and too timid to cut the mustard. In any event, I was certainly influenced by what I was trying to critique, sharing in some ways the confusion and mystification I was trying to dispel, and hence not able to articulate my concerns and objections with sufficient clarity to be effective. Therefore, as a contribution to clarifying CSCW’s program and thereby, I hope, help bringing it out of the current crisis, I will try to disentangle the issue of ‘plans and situated actions’ and then move on to the concepts of computation and computational regulation. 1. Suchman vs. cognitivism Suchman’s book is a sharp and generally well-articulated critique of ‘cognitive science’ as exemplified by the cognitivist tradition within cognitive psychology as well as by its intellectual counterparts in ‘artificial intelligence’ and the offspring Chapter 12: Frail foundations 387 of this in the form of ‘expert systems’ and ‘office automation’. The focal point of her critique is the concept of ‘plans’ as it was conceived of by cognitivist theorists. Her target was the view ‘that purposeful action is determined by plans’ and that this was considered ‘the correct model of the rational actor’, and what motivated her critique was the observation that this view was ‘being reified in the design of intelligent machines’ (Suchman, 1987, p. ix). In this regard, the formulation given by George Miller, Eugene Gelernter, and Karl Program in their book Plans and the Structure of Behavior (1960) is representative: ‘Any complete description of behavior should be adequate to serve as a set of instructions, that is, it should have the characteristics of a plan that could guide the action described. When we speak of a Plan in these pages, however, the term will refer to a hierarchy of instructions, and the capitalization will indicate that this special interpretation is intended. A Plan is any hierarchical process in the organism that can control the order in which a sequence of operations is to be performed. A Plan is, for an organism, essentially the same as a program for a computer.’ (Miller, et al., 1960, p. 16; quoted in Suchman, 1987, pp. 36 f.). Suchman summarized the basic tenets of the movement as follows: ‘The cognitivist strategy is to interject a mental operation between environmental stimulus and behavioral response: in essence, to relocate the causes of action from the environment that impinges upon the actor to processes, abstractable as computation, in the actor’s head. The first premise of cognitive science, therefore, is that people — or “cognizers” of any sort — act on the basis of symbolic representations: a kind of cognitive code, instantiated physically in the brain […]. The agreement among all participants in cognitive science and its affiliated disciplines, however, is that cognition is not just potentially like computation, it literally is computational.’ (Suchman, 1987, p. 9). In opposition to this view and drawing on various resources, principally anthropology and sociology (e.g., ethnomethodology and conversation analysis), Suchman developed a powerful argument based on the concept of ‘situated actions’: ‘By situated actions I mean simply actions taken in the context of particular, concrete circumstances’ (p. viii). Her key argument is that ‘however planned, purposeful actions are inevitably situated actions’ ‘because the circumstances of our actions are never fully anticipated and are continuously changing around us. As a consequence our actions, while systematic, are never planned in the strong sense that cognitive science would have it. Rather, plans are best viewed as a weak resource for what is primarily ad hoc activity. It is only when we are pressed to account for the rationality of our actions […] that we invoke the guidance of a plan. Stated in advance, plans are necessarily vague, insofar as they must accommodate the unforeseeable contingencies of particular situations. Reconstructed in retrospect, plans systematically filter out precisely the particularity of detail that characterizes situated actions, in favor of those aspects of the actions that can be seen to accord with the plan.’ (pp. viii f.) Whereas the course of rational action, on the cognitivist view, is causally determined by some, putative, preformed ‘plan’, ‘scheme’, etc., Suchman argued for an alternative account: ‘The alternative view is that plans are resources for situated action, but do not in any strong sense determine its course. While plans presuppose the embodied practices and changing circumstances 388 Cooperative Work and Coordinative Practices of situated action, the efficiency of plans as representations comes precisely from the fact that they do not represent those practices and circumstances in an of their concrete detail.’ (p. 52) This account is expressed in a slightly more specified manner by the end of the book: ‘For situated action […] the vagueness of plans is not a fault, but is ideally suited to the fact that the detail of intent and action must be contingent on the circumstantial and interactional particulars of actual situations’. ‘Like other essentially linguistic representations, plans are efficient formulations of situated actions. By abstracting uniformities across situations, plans allow us to bring past experience and projected outcomes to bear on our present actions. As efficient formulations, however, the significance of plans turns on their relation back to the unique circumstances and unarticulated practices of situated activities.’ (pp. 185 f.) The influence Suchman’s book has had is of course to a large extent due to this critique of cognitivism. In fact, for many researchers in the communities concerned with human factors of computing (HCI, PD, CSCW), the book had the effect of something like a declaration of independence. But it also, and not least, provided initial formulations of the practice-oriented research program for which her previous work on office procedures had provided one of the exemplars: ‘I have introduced the term situated action. That term underscores the view that every course of action depends in essential ways upon its material and social circumstances. Rather than attempting to abstract action away from its circumstances and represent it as a rational plan, the approach is to study how people use their circumstances to achieve intelligent action. Rather than build a theory of action out of a theory of plans, the aim is to investigate how people produce and find evidence for plans in the course of situated action. More generally, rather than subsume the details of action under the study of plans, plans are subsumed by the larger problem of situated action.’ (p. 50) Not only did Suchman offer an initial formulation of the practice–oriented research program; she did so in a way that from the very outset pointed to the concept of the embodiment of action and the materiality of work practices as foundational: ‘all activity, even the most analytic, is fundamentally concrete and embodied’ (p. viii); accordingly, the materiality of the work setting is not a liability but an asset to the practitioner: ‘the world is there to be consulted should we choose to do so’ (p. 47); ‘plans presuppose the embodied practices and changing circumstances of situated action’ (p. 52). In doing so, she emphasized and outlined critical aspects of the understanding of cooperative work that subsequently evolved in CSCW. However, Suchman’s book has also left an intellectual legacy that hampers CSCW with respect to addressing critical aspects of the real-world problems that it initially set out to address and which remain a domain it is uniquely equipped to address: the design and use of computational regulation devices as a means of dealing with the complexities of coordinative practices in cooperative work. Suchman of course did not deny that ‘plans’ are produced and used, nor did she say or imply that ‘plans’ are more or less useless. The problems with her account are far more subtle than that. In my attempt to unravel these problem, I will first show that Suchman, unwittingly and tacitly, accepted basic premises of the cognitivist position she is trying to dismantle. I will argue that she, because of this, was Chapter 12: Frail foundations 389 unable to dispose of with cognitivism’s confusion of normative and causal propositions. At the end she therefore wound up with an account that effectively reproduced cognitivism’s view on the nature of computational artifacts. 1.1. Suchman’s strategy First of all, what kind of argument is presented in Plans and Situated Actions? It has been argued that the focus of Suchman’s argument ‘is on the notion of plan as deployed in cognitive science (plans-according-to-cognitive-science)’ and that, therefore, much of her ‘argumentation does not concern “plans” as we might use them in ordinary affairs’ (Sharrock and Button, 2003, p. 259). It is certainly true that it is the cognitive science notion of ‘plans’ Suchman is critiquing in her book and that her propositions should be read in that context. However, this gallant reading is not supported by the text, and it has in fact been rejected by Suchman: ‘My aim was to take on both senses of “plan,” […] and to explore the differences between them’ (Suchman, 2003, p. 300). Indeed, Suchman’s research prior writing the book had focused on ‘“plans” as we might use them in ordinary affairs’, namely, organizational procedures, and in that work she did talk about ordinary ‘procedures’ in exactly the same terms as she talked about ‘plans’ in the book (cf. Suchman, 1982, 1983; Suchman and Wynn, 1984). This concordance justifiably led Agre to read the book as a book about ordinary plans (Agre, 1990). However, for unknown reasons, those earlier and concordant studies seem to have been forgotten by those who have later defended the book against its critics. Anyway, her exploration of the differences between the cognitivist and the ordinary sense of ‘plan’ was not the conceptual analysis that would have been required to do the job. If it had been that kind of argument she would have been trying to expose the deep conceptual confusion underlying cognitivism. But that she did not do. Cognitivism is a special version of what Ryle calls ‘the intellectualist legend’. This legend is characterized by having implicitly taking the assumed pattern of intellectual conduct (e.g., a cycle of theorizing, planning, acting, evaluating) as the paradigm of all intelligent conduct. The theorist thus explains intelligent action by ascribing an anterior ‘plan’ to the action, the latter being the execution of this occult ‘plan’. But as pointed out by Ryle, ‘Intelligent practice is not a stepchild of theory. On the contrary theorising is one practice amongst others and is itself intelligently or stupidly conducted’ (Ryle, 1949, p. 26). Preconceived plans, whether ascribed or avowed, can be as smart or as stupid as any action, planned or not. A tennis player who is playing without a preconceived plan, or who does not stop and contemplate her next move, is not necessarily playing mindlessly. We sometimes make the effort of developing plans for our actions and we sometimes postpone action until we have a plan, but we most often simply act and, thankfully, in doing so we normally act intelligently, competently, heedfully, etc. If anterior plans were a requisite for intelligent conduct, then the development of plans would in turn require anterior plans to be intelligent, and so on. ‘To put it quite 390 Cooperative Work and Coordinative Practices generally, the absurd assumption made by the intellectualist legend is this, that a performance of any sort inherits all its title to intelligence from some anterior internal operation of planning what to do’ (Ryle, 1949, p. 31). The cognitivist version of ‘the intellectualist legend’ is one that accounts for intelligent practice in terms of (postulated, occult) causal processes. As already shown, Miller et al. (1960) jumps from ‘description of behavior’ to ‘a set of instructions’ to ‘plan’ to organic ‘process’ to sequential ‘control’ to ‘computer program’. In the words of Stuart Shanker in his incisive critique of cognitivism, this ‘muddle’ is the ‘result of trying to transgress logical boundaries governing the employment of concepts lying in disparate grammatical systems’ (Shanker, 1987c, p. 73). In other words, the muddle is the result of an entire series of category mistakes in close order. I will develop this line of argument below. The point here is that Suchman does not deploy this kind of argument. Instead she counters the conceptual confusion of cognitivism by propounding what is basically an empirical argument, an ‘alternative account of human action’ (Suchman, 1987, p. x), drawing on observational studies and conceptualizations from social science. This mismatch, an empirical argument against conceptual confusion, is a significant source of ambiguity. One more observation on the kind of argument that is presented in Plans and Situated Actions is required. Instead of a conceptual critique of cognitivism, Suchman mobilized an array of social science accounts of human action and interaction: ethnomethodology, conversation analysis, studies of instructions, etc. – as if to say, ‘Look, the cognitivist account is not realistic. The sociological account offers a richer picture’. Paradoxically, however, if this was indeed her aim, the examples of ‘plans and situated actions’ she offered and discussed are far from representative of ordinary plans in ordinary affairs. Here, in a critique of cognitivism premised on presenting an alternative account, one would have expected that the rich multiplicity would have been demonstrated. For does it make any sense at all to talk about ‘plans’ in abstract generality, as if it is a genuine concept? Should we not say of the word ‘plan’ what Wittgenstein says of the word ‘to think’: ‘It is not to be expected of this word that it should have a unified employment; we should rather expect the opposite’ (Wittgenstein, 1945-48, § 112). The concept of plan is certainly multifarious in its uses. We talk about the floor plan of a building, CAD plans, production plans or schedules, maintenance plans, project plans, cancer treatment plans, and so on. What these uses have in common is the central role of some artifact. In fact, the English word ‘plan’ is derived from French ‘plan’ (a ground plan or map, as in ‘plan de ville’, from the Latin ‘planum’, a flat surface). The notion is of a drawing on a flat surface, a standard of correctness that can be used as instruction for action and guide in action and that can be inspected and consulted in case of doubt, used as proof in case of dispute, and so on. We certainly also talk about plans in a derived sense in other contexts of ordinary discourse. We use the term, for instance, to claim not only intent but considered intent: ‘I plan to leave early so that I can meet you by noon’, meaning some- Chapter 12: Frail foundations 391 thing like ‘I have indeed been considering when to leave and decided to depart early so that I’m certain to arrive at our rendezvous by noon.’ By using the term ‘plan’ in that way in such a context one is declaring not only commitment but also that one has given one’s promise some serious thought. One might elaborate by saying: ‘I plan to depart from point A at 9:00, which means that I’ll be at point B at noon: As promised.’ Such avowals of considered commitment derive their force from using the term ‘plan’, with its received connotations of publicly visible inscription, metaphorically: my commitment is as firm as if it was on public record. Now, Suchman also ascribed ‘plans’ to situations where such avowals may not or need not to have been uttered, for example to what goes on prior to white water canoeing: ‘in planning to run a series of rapids in a canoe, one is very likely to sit for a while above the falls and plan one’s descent. The plan might go something like “I’ll get as far over to the left as possible, try to make it between those two large rocks, then backferry hard to the right to make it around that next bunch.”’ (p. 52) It should be noticed, in passing, that this example, possibly due to its brevity, is ambiguous. Somebody is contemplating how to approach a line of action and we are told that the plan may require a ‘great deal of deliberation, discussion, simulation, and reconstruction’. We are not told if the plan has any public status, whether the articulation of the plan amounts to an avowal of commitment. That is, what would happen if the planner or her co-canoeist disregards the plan they have committed to? If one of them breaks the plan and they end up wet, cold, and bruised, could the other not then object, ‘But we agreed to backferry hard to the right, not to the left’. As it stands, the example reads as if the ‘plan’ and its possible ‘abandonment’ is somehow of no consequence to either of them or to others. In what sense is it a plan then? Suchman also referred to ‘plans’ belonging to practices that, by contrast, involve the use of inscribed artifacts, such as a traveller (a lone traveller!) using a map to find his way (p. 189). But a map in the hands of a traveller belongs to an entirely different kind of practice than someone’s contemplating the course of a canoe trip. And it can, in turn, be used in quite different ways, as a representation of the geography and what it may offer to a traveller with time to spare or as a representation of the trajectory one wants to or, indeed, is obliged to follow. And, again, ‘plans’ incorporated in the help system of a photocopier, the use of which Suchman analyzed in the book, belong to practices of yet another kind. In this case the putative ‘plans’ are computational artifacts that are supposed to regulate peoples’ use of the photocopier. That is, as opposed to the map in the hands of the bored traveller, this artifact has the capacity to execute controlled state changes which may physically prevent a user from (or enable him in) doing this or that in certain circumstances. Whatever merit the particular design may have or lack, a practice in which a causal mechanism (a computational artifact) is used as part of a technique of planning is certainly quite different from a practice in which the artifact of planning is static, which again is quite different from a practice in 392 Cooperative Work and Coordinative Practices which one has promised, on one’s honor, to perform a task in a certain way, which, finally, surely is different from a practice in which a person is pondering the best course of canoeing. These various practices may have something in common (beyond the noun ‘plan’ we tend to use when referring to these practices) but nothing that would make ‘plans’ in all fuzzy generality a researchable phenomenon. What is most remarkable, however, is that any discussion of ordinary plans — time tables, production schedules, project plans, clinical protocols — is absent from the analysis. It would not be surprising were such real-world plans systematically omitted from cognitivist discourse,104 but their absence in the context of a call for ethnographic studies of actual socially situated action is very remarkable indeed. Their absence is all the more remarkable in light of the fact that Suchman’s early work exactly focused on practices in which such ordinary plans loom large (e.g., Suchman, 1983). 1.2. Counter-cognitivism On a decisive issue Suchman’s critique of cognitivism is very clear and firm. She stated — emphatically and repeatedly — that ‘our actions, while systematic, are never planned in the strong sense that cognitive science would have it’ (p. ix); that ‘plans are resources for situated action, but do not in any strong sense determine its course’ (p. 52), and so on. By the expression ‘strong sense’ she obviously meant ‘causal sense’. That is, she took issue with the basic cognitivist proposition that rational action is to be explained by reference to some (obscure) causal ‘process’. However, she did not realize that in ordinary language the concept ‘plan’ (‘following a plan’, ‘agreeing to a plan’, ‘violating a plan’) is a concept of normative behavior. Hence an expression such as ‘planning in the strong [causal] sense’ is unintelligible, as it suggests the very possibility of a causal determination of action by a plan. Plans are normative constructs; that is, they provide criteria for whether or not a particular action is correctly executed. On the other hand, if Suchman was using the word ‘plan’ in the sense underlying cognitivism’s metaphysical language (material processes that somehow maintain other material processes in a certain state) then her key propositions in turn loose their sense. What might it, for instance, then mean to say that, ‘as essentially linguistic representations, plans are efficient formulations of situated actions’ (p. 186)? The problem is, of course, that if one uses the word ‘plan’ in a contrived sense, transposed from a quite different domain of discourse, then one has lost the 104 Miller et al. actually do refer to ordinary plans: ‘A public plan exists whenever a group of people try to cooperate to attain a result that they would not be willing or able to achieve alone. Each member takes upon himself the performance of some fragment of the public plan and incorporates that fragment into his individual, personal Plans’ (Miller, et al., 1960, p. 98). However, they obviously do not notice that they are here implying ‘plans’ in a normative sense. Chapter 12: Frail foundations 393 ability to account for ordinary plans in our everyday life. (At the same time one has introduced a source of confusion in computer science as well). Thus, when Suchman talked about ‘plans’, it is generally not clear if the word is to be read as ‘plans-according-to-cognitive-science’ or as ‘“plans” as we might use them in ordinary affairs’ (Sharrock and Button, 2003, p. 259). As noted above, this was deliberate on Suchman’s part. This is more than a source of ambiguity, however. Suchman was herself obviously not aware of this rather important distinction. The normative status of plans is completely absent from her account. Cognitivism’s foundations were thus left effectively untouched and intact. Having from the outset accepted the cognitivist concept of ‘plan’ as meaningful, albeit empirically objectionable, and having thus conceded the high ground, how could Suchman then possibly mount anything like an attack on cognitivism? She did so by deploying what could be dubbed a strategy of containment. That is, she tried to keep the cognitivist peril at bay by neutralizing its account of rational action and the pernicious implications that flow from it, by insisting that ‘plans are best viewed as a weak resource for what is primarily ad hoc activity’, etc. The intellectual costs of this strategy are very high indeed, however. To protect the phenomenon of real-world human action — socially organized, materially situated, embodied, intentional action — from being reduced to a mere epiphenomenon of some hidden causal process, Suchman introduced a categorical gap between plan and situated action. She did that, for instance, when she stated that ‘situated action turns on local interactions between the actor and contingencies that, while they are made accountable to a plan, remain essentially outside of the plan’s scope’ (pp. 188 f.). The proposition is somewhat ambiguous in as much as it is not entirely clear what is meant by the term ‘scope’, but the expression ‘essentially outside’ would normally serve to indicate that the concepts of ‘situated action’ and ‘plan’ are conceptually unrelated, i.e., that they are unrelated by definition, and that any relation therefore would be contingent. This reading is confirmed by statements throughout the book in which ‘plans’ are granted a role prior to and subsequent to action, but not in action. For example, and as shown above, Suchman asserted that ‘It is only when we are pressed to account for the rationality of our actions, given the biases of European culture, that we invoke the guidance of a plan’ (p. ix). In a similar vein she posited that ‘plans are a constituent of practical action, but they are constituent as an artifact of our reasoning about action, not as the generative mechanism of action’ (p. 39). The reading I am suggesting, that Suchman construed a categorical separation of ‘plans’ and ‘situated actions’, is further confirmed by her discussion of the canoe example mentioned in passing above: ‘A great deal of deliberation, discussion, simulation, and reconstruction may go into such a plan. But, however detailed, the plan stops short of the actual business of getting your canoe through the falls. When it really comes down to the details of responding to currents and handling a canoe, you effectively abandon the plan and fall back on whatever embodied skills are available to you. The purpose of the plan in this case is not to get your canoe through the rapids, but rather to orient you in such a way that you can obtain the best possible position from which to use those embodied skills on which, in the final analysis, your success depends.’ (p. 52) 394 Cooperative Work and Coordinative Practices That is, on Suchman’s view, when one starts acting, one ‘effectively abandons’ the plan; its only purpose now is to ‘orient’ one so as to be able to improvise smartly. The same picture of plans or ‘abstract representations of situations and of actions’ was painted in the conclusions of the book: ‘The foundation of actions […] is not plans, but local interactions with our environment, more and less informed by reference to abstract representations of situations and of actions, and more and less available to representation themselves. The function of abstract representations is not to serve as specifications for the local interactions, but rather to orient or position us in a way that will allow us, through local interactions, to exploit some contingencies of our environment, and to avoid others.’ (p. 188). Again action was portrayed as essentially ad hoc action, ‘local interactions’, and again the role of ‘plans’ was seen as merely that of affording improvisation: ‘orient or position us in a way that will allow us, through local interactions, to exploit some contingencies of our environment, and to avoid others’. Now, Suchman did state, quite clearly, that ‘the essential nature of action, however planned or unplanned, is situated’ (p. x). I can of course have no problem with that statement. All action, planned or improvised, is situated, i.e., contingent, embodied, materially contextualized, etc. But it is important to realize that this is not a substantive or empirical proposition but a logico-grammatical one. It simply states that action and context are internally related concepts. But Suchman tended to forget this. As already shown, ‘situated actions’ were often characterized as ‘essentially ad hoc’ (e.g., pp. ix, 48, 61, 78), which is plain nonsense. Action is essentially (!) situated, and some actions are ad hoc, while other actions certainly are not ad hoc. The action of executing a plan is, being an action, situated; but it is not, again by definition, ad hoc. In sum, it is internal to the concept of action that action is situated, contingent. To say that action is essentially ad hoc or that plans flounder on the contingent nature of action is deeply confused. Some action is characterized by being overwhelmingly spontaneous, unpremeditated, ad hoc, improvised, etc. Other action is planned in the sense that there is an obligation to execute the action in certain ways: steps to be taken in a certain sequence, by certain actors, at certain times, by using certain resources, etc. Plans do not cause action to take a particular course, for they cannot cause anything. Just like rules, conventions, notations, etc., plans are normative constructs of our common practices. Chapter 12: Frail foundations 395 1.3. Transcendental judgments ‘We understand what it means to set a pocket-watch to the exact time, or to regulate it to be exact. But what if it were asked: Is this exactness ideal exactness? […] No single ideal of exactness has been envisaged; we do not know what we are to make of this idea’ (Wittgenstein, 1945-46, § 88) The costs of Suchman’s strategy of containing cognitivism are not limited to the categorial separation of ‘plans’ and ‘actions’ and to the ‘abandonment’ of plans at a desolate place outside of situated action. As we have seen, the cognitivist notion of ‘plan’ is predicated on a notion of ‘a complete description of behavior’ (Miller, et al., 1960). Within a specific practice the notion of a complete description or a completely specified plan, instruction, recipe, etc. of course make sense, in as much as there are criteria for what can be taken to be a completely specified plan, etc. That is, the plan, the instruction, etc. is complete when it for an ordinary practitioner is unproblematic to follow it, apply it, use it, etc. However, the cognitivist notion of a ‘a complete description of behavior’ presumes completeness in the abstract, irrespective of any specific practice. The cognitivist notion of a ‘complete’ plan is therefore as absurd as the fantastic story of the ‘perfect map’ that Borges has given us. The story is presented as a text fragment entitled ‘Of Exactitude in Science’ that Borges, playfully, pretends to have found in an old book, ‘Travels of Praiseworthy Men (1658) by J. A. Suarez Miranda’: ‘… In that Empire, the craft of Cartography attained such Perfection that the Map of a Single province covered the space of an entire City, and the Map of the Empire itself an entire Province. In the course of Time, these Extensive maps were found somehow wanting, and so the College of Cartographers evolved a Map of the Empire that was of the same Scale as the Empire and that coincided with it point for point. Less attentive to the Study of Cartography, succeeding Generations came to judge a map of such Magnitude cumbersome, and, not without Irreverence, they abandoned it to the Rigours of sun and Rain. In the western Deserts, tattered Fragments of the Map are still to be found, Sheltering an occasional Beast or beggar; in the whole Nation, no other relic is left of the Discipline of Geography.’ (Borges, 1946, p. 141) Here the absurdity is obvious enough. But the very idea that a ‘complete description’ could be given outside of a practice in which the criterion for completeness is given is equally absurd. Suchman was at pains to demonstrate that the cognitivist completeness notion is untenable, which it certainly is. The problem with Suchman’s strategy is that she engaged in an overwhelmingly empirical argument to demonstrate that the metaphysical assumptions of cognitivism are factually groundless. But they are not groundless; they are meaningless. Her strategy therefore, unwittingly, reproduced the metaphysics of cognitive science, albeit in the inverse. This shows in several ways. 396 Cooperative Work and Coordinative Practices Suchman repeatedly stated that plans are inherently ‘vague’ compared with action: ‘Stated in advance, plans are necessarily vague, insofar as they must accommodate the unforeseeable contingencies of particular situations’ (p. ix). Similarly, she stated that ‘plans are inherently vague’ (pp. 38, 185) or refers to the ‘representational vagueness’ of plans (p. 185). But could action be anything but ‘situated’ in the sense used by Suchman: contextual, embodied, etc.? Would ‘doing x’ not be categorically different from ‘thinking about doing x’ (or ‘describing doing x’ or ‘planning doing x’)? But of course it would! This is as trivial as saying that there is a difference between the landscape and the map of the landscape. How does one compare the ‘vagueness’ or ‘completeness’ of ‘plans’ vis-à-vis ‘actions’? Is there a metric out there one can employ? Surely not. But what does it mean, then, that ‘plans are inherently vague’? When Suchman was making these statements she was falling back into the metaphysical framework of the cognitivist tradition she was otherwise trying to demolish. The cognitivists’ transcendental use of the term ‘completeness’ and Suchman’s equally transcendental use of the opposite term ‘vagueness’ are both examples of the kind of metaphysical smoke that is produced when language is idling. The notion of ‘vague’ plans, ‘incomplete’ plans, etc. presupposes the logical possibility of distinct plans, complete plans, etc. The terms ‘vague’ and ‘complete’ are used as characterizations of specific plans with respect to specific practices. According to which criteria can a plan be said to be ‘vague’? Are there criteria of ‘vagueness’ or specificity or adequacy outside of the particular practice? Suchman pointed out herself that ‘While plans can be elaborated indefinitely, they elaborate actions just to the level that elaboration is useful; they are vague with respect to the details of action precisely at the level at which makes sense to forego abstract representation, and rely on the availability of a particular, embodied response’ (p. 188). True: provided one has indefinite time and resources, then plans can be elaborated indefinitely. But if the criterion of a plan’s appropriate level of elaboration is a practical one, then surely some plans are vague (with respect to the criteria internal to the practice in question) and others not. In fact, we would hardly make plans if they were not, generally, suitably specific and complete for our practical purposes. The same metaphysical form of discourse is in evidence when Suchman stated that the ‘circumstances of our actions […] are continuously changing around us’ (p. viii). Sure, there is a sense in which we can say that the world changes continually: electrons jump about, molecules form and dissolve, cells reproduce and decay, etc. But the notion of the ‘circumstances of our actions’ refers to the circumstances that are of practical significance, those that to a practitioner make a difference. And in that sense the ‘circumstances of our actions’ may or may not change. Hence, we can unproblematically talk about doing the same thing under the same circumstances. That is, in her honorable effort to turn the table on the cognitivism’s realist notion of ‘complete description’ in the abstract, Suchman fell back Chapter 12: Frail foundations 397 into the trite nominalist notion that ‘everything is unique’: one cannot jump into the same river twice, nothing stays the same, all action is therefore ad hoc. 105 The source of these confused ideas is, I suggest, the cognitivist notion that plans are ‘descriptions’, ‘representations’, etc. On the cognitivist account, ‘plans’ are hypotheses derived inductively that (in some unintelligible way) are operating causally. To this Suchman countered with the classical objection that inductively derived hypotheses (and theories) are underdetermined with respect to manifold reality and that a given hypothesis therefore is only one possible interpretation out of many. Hence, I presume, the confusing talk about essential vagueness of ‘plans’. In sum, Suchman was trying to demolish the cognitivist program but she did not, from the very outset, dissolve cognitivism’s obscure notion of ‘plans’. That is, she accepted to oppose the cognitivist notion of ‘plans’ on the very battle field chosen and defined by cognitivism. The cognitivist notion of ‘plans’, phantasmal objects that supposedly determine rational action in a causal sense, is a meaningless construct and would have to be demolished as such. By accepting this construct as having sense, Suchman also, unwittingly and against everything she otherwise stood for, engaged in a metaphysical discourse and got trapped as a fly in a bottle. For CSCW the immediate problem with the metaphysics underlying cognitivism as well as that retained in Suchman’s reversal of cognitivism is that this form of theorizing tends to make us blind to the multiplicity of practices and hence to phenomena that are crucially important to us. Consequently, although Suchman’s book was justly received enthusiastically as portending a liberation from the cognitivist dogma and from its stifling effects on socio-technical research, and while it offered vital initial formulations of the practice-oriented program of CSCW, it also, at the same time, contributed to stifling that program by reproducing the metaphysical form of reasoning characteristic of cognitivism. 105 Suchman has recently, in response to criticisms of propositions such as ‘situated actions are essentially ad hoc’, made some rather guarded comments: ‘I see my choice of the term ad hoc here as an unfortunate one, particularly in light of subsequent readings of the text. The problem lies in the term’s common connotations of things done anew, or narrowly, without reference to historically constituted or broader concerns. Perhaps a better way of phrasing this would be to say that situated actions are always, and irremediably, contingent on specific, unfolding circumstances that are themselves substantially constituted by those same actions. This is the case however much actions may also be informed by prescriptive representations, past experience, future considerations, received identities, entrenched social relations, established procedures, built environments, material constraints and the like. To be rendered effective the significance and relevance of any of those must be reiterated, or transformed, in relation to what is happening just here and just now.’ (Suchman, 2007, p. 27, n. 4). — It first of all needs to be said that the ‘choice of the term ad hoc’ is not simply ambiguous but confusing in as much as ‘ad hoc’, as any dictionary will confirm, has a stable meaning, namely, something done ‘for the particular end or purpose at hand and without reference to wider application or employment’, in contrast to ‘planned’ or ‘coordinated’. Anyway, removing the term from the text does not alter anything. Should the phrase ‘contingent on specific, unfolding circumstances’ be read as a logicogrammatical one or an empirical one? In case the former reading is intended, the clause simply states that action is irremediably situated or that planning is not acting. In the latter reading, the question remains: ‘irremediably contingent’ according to which criteria? Does it, on this account, even make sense to talk of an action as an identifiable course of conduct? 398 Cooperative Work and Coordinative Practices 1.4. Regularity and normativity ‘A rule stands there like a signpost. — Does the signpost leave no doubt about the way I have to go? Does it show which direction I am to take when I have passed it, whether along the road or the footpath or crosscountry? But where does it say which way I am to follow it; whether in the direction of its finger or (for example) in the opposite one?’ ‘The signpost is in order — if, under normal circumstances, it fulfils its purpose.’ (Wittgenstein, 1945-46, §§ 85, 87) The extent to which Suchman in her book stayed within the cognitivist framework is particularly clear in her understanding of the term ‘plan’. For cognitivism, for which persons are mere carriers of ‘plans’, a ‘plan’ is a description of action. On this view, the normative sense of plans, as prescribed or agreed–to course of action, is not considered at all. Suchman conceived of plans in exactly the same way: ‘in our everyday action descriptions we do not normally distinguish between accounts of action provided before and after the fact, and action’s actual course’ (p. 38 f.). ‘Like all action descriptions, instructions necessarily rely upon an implicit et cetera clause in order to be called complete’ (p. 62). ‘The general task in following instructions is to bring canonical descriptions of objects and actions to bear on the actual objects and embodied actions that the instructions describe’ (p. 101). This is again remarkable. As already noted, the term ‘plan’ is generally used in ways comparable to those of the term ‘rule’. We use the term ‘rule’ both descriptively, to indicate regularity (‘As a rule, I clock out when the bars open’), or as a criterion of correct conduct (‘The house rule says no fighting and no spitting on the floor!’). Similarly, we certainly sometimes use the term ‘plan’ as an ‘action description’: ‘He seemed to be working according to a plan’, or ‘He acted in conformance with Plan B’; but we also use the term to refer to normative constructs: ‘This is our plan…’ or ‘According to the production plan, we have to be finished today.’ How can we talk and reason sensibly about rules and plans and ad hoc activities and about the role of ‘formalisms’ in work practices? To extract ourselves from the quagmire of cognitivism, which Suchman set out to do but did not succeed in doing, we should first of all heed some cardinal distinctions of which Wittgenstein has reminded us.106 A major source of confusion is many social scientists’ apparently stubborn refusal to distinguish between, on one hand, mere regularity of behavior, and on the other hand, following a rule. 106 The following discussion is, of course, primarily based on Wittgenstein’s famous analysis in Philosophical Investigations (1945-46, esp. §§ 82-87 and 185-242) but also his Remarks on the Foundations of Mathematics (1937-44). For excellent but not entirely mutually congruent commentaries on this subtle analysis, cf. the works of Winch (1958), Pitkin (1972), von Savigny (1974, 1994), Baker and Hacker (1984, 2009), Malcolm (1986, 1995), Hacker (1989, 1990), and Williams (1999). Chapter 12: Frail foundations 399 Mere regularity: that is, people’s exhibiting observable and reasonably predictable patterns of behavior, or their acting in observable conformity with a rule. What many sociologists and psychologists try to tease out by trying to detect correlations in behavior (e.g., differing suicide rates) is this: mere regularity. Such patterns may be important ‘sociological facts’, but they may also be of the same stuff as the notorious channels on Mars. When we talk about rule following, by contrast, we talk about something entirely different: practices that not only involve observable regularity of conduct but also the ability of actors to explain, justify, sanction, reprimand, etc. actions with reference to rules, and often also the ability to teach rules, formulate rules, debate rules, etc. In his Remarks on the Foundations of Mathematics, Wittgenstein gives an wonderfully straightforward illustration of this point: ‘Let us consider very simple rules. Let the expression be a figure, say this one: ⎢— —⎥ and one follows the rule by drawing a straight sequence of such figures (perhaps as an ornament). ⎢— —⎥ ⎢— —⎥ ⎢— —⎥ ⎢— —⎥ ⎢— —⎥ Under what circumstances should we say: some gives a rule by writing down such a figure? Under what circumstances: someone is following this rule when he draws this sequence? It is difficult to describe this. If one of a pair of chimpanzees once scratched the figure ⎢— —⎥ in the earth and thereupon the other the series ⎢— —⎥ ⎢— —⎥ etc., the first would not have given a rule nor would the other be following it, whatever else went on at the same time in the mind of the two of them. If however there were observed, e.g., the phenomenon of a kind of instruction, of shewing how and of imitation, of lucky and misfiring attempts, of reward and punishment and the like; if at length the one who had been so trained put figures which he had never seen before one after another in sequence as in the first example, then we should probably say that the one chimpanzee was writing rules down, and the other was following them.’ (Wittgenstein, 1937-44, VI §42) The difference between ‘regularity’ and ‘rule following’ is a categorial one. The formulation of a rule is not an empirical proposition, whereas the formulation of a regularity is; the formulation of a rule is a normative one, it provides criteria for what is correct and what is not, what is right and what is wrong. One may, for example, observe that people of a certain age are disproportionately represented among those who commit suicide; this would be an observation of a regularity. But making this observation is obviously not the expression of a rule. If I said to a particular person that he had tried to commit suicide at the wrong age, I would be most certainly be regarded as demented, and rightly so. This should be clear enough, or so I should like to think. Let us therefore move on to the more tricky issue of the relationship between rule and action. What may make Wittgenstein’s observation on our concept of ‘following a rule’ particularly hard to accept by sociologists and other social scientists, is the strong inclination in social theory to conceive of rational action as necessarily involving some kind of ‘interpretation’ of rules, instructions, precedents, etc. on the one hand and ‘interpretation’ of the situation on the other. This is the intellectualist legend at work: the actor is portrayed in the image of the scholar bent over an 400 Cooperative Work and Coordinative Practices ancient text fragment trying to develop a version that is both loyal to the original and at the same time understandable to the modern reader. Wittgenstein reminds us, however, not to confound following a rule and interpreting a rule. To do so he takes the reader of his Philosophical Investigations through a lengthy reductio ad absurdum, demonstrating that a course of action that exhibits regularity can be made out to conform with multiple and mutually contradictory rule formulations. After having done that, Wittgenstein stops the reductio and lets his fictional interlocutor ask: ‘But how can a rule teach me what I have to do at this point? After all, whatever I do can, on some interpretation, be made comparable with the rule’. Wittgenstein replies: ‘No, that’s not what we should say. Rather: every interpretation hangs in the air together with what it interprets, and cannot give it any support. Interpretations by themselves do not determine meaning.’ (Wittgenstein, 1945-46, § 198). The operative word here is interpretation. A few sections later, Wittgenstein emphasizes this point: ‘This was our paradox: no course of action could be determined by a rule, because every course of action can be brought into accord with the rule. The answer was: if every course of action can be brought into accord with the rule, then it can also be brought into conflict with it. And so there would be neither accord nor conflict here. That there is a misunderstanding here is shown by the mere fact that in this chain of reasoning we place one interpretation behind another, as if each one contend us at least for a moment, until we thought of yet another lying behind it. For what we thereby show is that there is a way of grasping a rule which is not an interpretation, but which, from case to case of application, is exhibited in what we call “following the rule” and “going against it”. That’s why there is an inclination to say: every action according to a rule is an interpretation. But one should to speak of interpretation only when one expression of a rule is substituted for another.’ (Wittgenstein, 1945-46, § 201) In following a rule there is no space for interpretation: ‘an interpretation gets us no closer to an application than we were before. It is merely an alternative formulation of the rule, another expression in the symbolism which paraphrases the initial one’ (Baker and Hacker, 2009, p. 92). To be sure, a course of action sometimes involves interpretation, but it does not necessarily do so. As Wittgenstein puts it in Zettel, ‘an interpretation is something that is given in signs. It is this interpretation as opposed to a different one (running differently)’ (Wittgenstein, 1945-48, § 229). That is, we talk of ‘interpretation’ when referring to the substitution of one linguistic construct (rule formulation, instruction, command, statement, representation, etc.) by another and, supposedly, more useful construct. In following a rule, the rule is not ‘interpreted’ or the like; it is simply applied or enacted, because that is what following a rule means. Understanding a rule means that I can apply it without engaging in interpreting the rule. As succinctly summarized by Baker and Hacker, ‘to grasp a rule is to understand it, and understanding a rule is not an act but an ability manifested in following the rule’ (Baker and Hacker, 2009, p. 96). And Wittgenstein again: ‘What happens is not that this symbol cannot be further interpreted, but: I do no Chapter 12: Frail foundations 401 interpreting. I do not interpret because I feel at home in the present picture. When I interpret, I step from one level of thought to another’ (Wittgenstein, 1945-48, § 234). In fact, action normally does not involve interpretation. In our ordinary work practices we do not normally engage in interpretation work whenever we follow instructions or execute plans. We sometimes have to, of course, but generally we do not. That is the whole point of the concept of ‘the natural attitude’. Interpretation is required when doubt is a practical issue, and doubt needs grounds too (Wittgenstein, 1949-51, § 124). Endless doubt is impossible: ‘If you tried to doubt everything you would not get as far as doubting anything. The game of doubting itself presupposes certainty’ (§ 115). We interpret when it is conceivable to us that we could be wrong, e.g., when we are uncertain about the meaning of a rule formulation or do not yet fully understand it. At the end of his discussion of the notion of following a rule in the Philosophical Investigations, Wittgenstein lets his interlocutor ask: ‘How am I able to obey a rule?’ To which Wittgenstein replies: ‘If this is not a question about causes, then it is about the justification for my acting in this way in complying with the rule. Once I have exhausted the justifications, I have reached bedrock, and my spade is turned. Then I am inclined to say: “This is simply what I do.”’ (Wittgenstein, 1945-46, § 217) Two paragraphs later, Wittgenstein wraps up this line of argument by saying: ‘When I follow the rule, I do not choose. I follow the rule blindly.’ (§ 219) If read out of context, the phrase ‘I follow the rule blindly’ can be misunderstood as suggesting that normative behavior is irrational, or non-rational: ‘But in context it signifies not the blindness of ignorance, but the blindness of certitude. I know exactly what to do. I do not chose, after reflection and deliberation, I just ACT — in accord with the rule’ (Baker and Hacker, 1984, p. 84). In other words, what is meant is not that the actor in following the rule proceeds mindlessly but that he or she goes on as a matter of course. What is meant by saying ‘I follow the rule blindly’ is made perfectly clear in the Remarks on the Foundations of Mathematics: ‘One follows the rule mechanically. Hence one compares it with a mechanism. “Mechanical” — that means: without thinking. But entirely without thinking? Without reflecting.’ (Wittgenstein, 1937-44, VII §61). When following the rule, that is, the actor proceeds as ‘a matter of course’ (Wittgenstein, 1945-46, § 238), for doubt is not an option. The rule ‘always tells us the same, and we do what it tells us’: ‘One does not feel that one has always got to wait upon the nod (the prompt) of the rule. On the contrary, we are not on tenterhooks about what it will tell us next, but it always tells us the same, and we do what it tells us.’ (Wittgenstein, 1945-46, § 223). The rule ‘is my final court of appeal for the way I’m to go’. On Wittgenstein’s account, then, the concept of ‘rule’ should be understood in its ‘internal’ or ‘logico-grammatical’ relations to 402 Cooperative Work and Coordinative Practices the concept of ‘practice’ and thereby to the concept of ‘techniques’ (Baker and Hacker, 2009, pp. 140-145). Understanding a rule means possessing the ability to do certain things correctly and is manifested in following a rule and mastering the appropriate techniques. If the question ‘How am I able to obey a rule?’ is about causes, however, then the answer is simply, in Wittgenstein’s words, that ‘We are trained to do so; we react to an order in a particular way’ (Wittgenstein, 1945-46, § 206). That is, ‘there is no explanation of our ability to follow rules — other than the pedestrian but true explanation that we received a certain training.’ (Malcolm, 1986, p. 180).107 For Wittgenstein, then, to follow a rule is a practice. A ‘person goes by a signpost only in so far as there is an established [ständigen] usage, a custom’ (§ 198). The rule, in contrast to the various spoken or written expressions and representations of the rule, does not exist independently of the action, as some mysterious mental entity. But nor does it make sense to think of rule following as something only one person could do only once in his or her life. ‘It is not possible that there should have been only one occasion on which only one person followed a rule. It is not possible that there should have been only one occasion on which a report was made, an order given or understood, and so on. — To follow a rule, to make a report, to give an order, to play a game of chess, are customs (usages, institutions)’ (Wittgenstein, 1945-46, § 199). To follow a rule means mastering a technique (§ 199) and is thus ‘a practice’ (§ 202). The concept of ‘practice’, then, should not be conceived of as mere conduct or behavior, nor as incessant improvisation or ‘irremediably contingent’ action. A practice is constituted by a rule (or an array of rules) that provides the standard of correct or incorrect conduct. It is the rule (or array of rules) that identifies a course of action as an instance of this practice, as opposed to an instance of another practice. In the words of Peter Winch, ‘what the sociologist is studying, as well as his study of it, is a human activity and is therefore carried on according to rules. And it is these rules, rather than those which govern the sociologist’s investigation, which specify what is to count as doing “the same kind of thing” in relation to that kind of activity’ (Winch, 1958, p. 87, emphasis deleted). The identity and integrity of rules and practices over time are themselves the result of practitioners’ ‘reflective’ efforts of instructing and teaching, of commanding and correcting, of emulating and practicing, of correcting own transgressions and asking for guidance, and of contemplating and negotiating new ways of doing things. Practices are upheld. 107 Meredith Williams has an interesting discussion of the key issue of learning in normative behavior (M. Williams, 1999, chapter 7). Chapter 12: Frail foundations 403 2. Work and interpretation work Why would Suchman be conceptually blind for the role of plans in action, that is, of the normative character of plans? The reason seems to be that she interposed interpretation between the plan and the action. Thus, to act the actor must first interpret the situation and the plan with respect to the situation and only then act. She seemed to believe that she was following Garfinkel in this, but that would be a misrepresentation of Garfinkel. Garfinkel is quoted (ibid., p. 62) for this observation: ‘To treat instructions as though ad hoc features in their use was a nuisance, or to treat their presence as grounds for complaining about the incompleteness of instructions, is very much like complaining that if the walls of a building were gotten out of the way, one could see better what was keeping the roof up.’ (Garfinkel, 1967, p. 22) It is a wittily put but carelessly general observation, in that the term ‘ad hoc’ of course presumes the logical possibility that instruction can be complete. The criterion of the completeness and incompleteness of instructions is internal to the particular practice. However, Suchman then elaborated: ‘Like all action descriptions, instructions necessarily rely upon an implicit et cetera clause in order to be called complete. The project of instruction-writing is ill conceived, therefore, if its goal is the production of exhaustive action descriptions that can guarantee a particular interpretation. What “keeps the roof up” in the case of instructions for action is not only the instructions as such, but their interpretation in use. And the latter has all of the ad hoc and uncertain properties that characterize every occasion of the situated use of language.’ (Suchman, 1987, p. 62) Suchman was here, again, making the argument that the map is not complete compared to the terrain and that its use therefore requires ‘interpretation’ and, accordingly, have ‘all of the ad hoc and uncertain properties that characterize every occasion of the situated use of language’. There is no reason to reiterate why this was confused. My point here is that she, explicitly, interposed interpretation as a necessary intermediary between instruction and action. (For a parallel critique, cf. Sharrock and Button, 2003). 2.1. Garfinkel (mis)interpreted It is highly relevant and instructive to note the nature of the case Garfinkel is referring to in the above quote. The study, conducted in the early 1960s by Garfinkel in collaboration with Egon Bittner was concerned with selection activities at an outpatient psychiatric clinic: ‘By what criteria were applicants selected for treatment?’. Their sources of information were the clinical records. The most important of these were intake application forms and the various contents of case folders. They take care to point out that clinical folders contain records that are generated by the activities of clinical personnel and that ‘almost all folder contents, as sources of data for our study, were the results of self-reporting procedures’ (Garfinkel, 1967, pp. 186 f.). 404 Cooperative Work and Coordinative Practices Altogether 1,582 clinic folders were examined by two graduate students of sociology who were tasked with extracting information and fill in a ‘coding sheet’. In doing this, the coders were permitted to make inferences and encouraged to undertake ‘diligent search’. Nonetheless, they were unable to obtain answers to quite many of the items in the coding sheet. For about half of the items dealing with clinical issues, the coders only got information from between 0 and 30 percent of the cases. Garfinkel and Bittner’s thorough account of the reasons for this rather dismal performance is a most informative discussion of the methodological problems that arise in studies that depend on secondary use of clinical records produced for internal use in the clinical setting. The gist of it is this: ‘We came to think of the troubles with records as “normal, natural” troubles. […] “Normal, natural troubles” are troubles that occur because clinic persons, as self-reporters, actively seek to act in compliance with rules of the clinic’s operating procedures that for them and from their point of view are more or less taken for granted as right ways of doing things. […] The troubles we speak of are those that any investigator — outsider or insider — will encounter if he consults the files in order to answer questions that depart in theoretical or practical import from organizationally relevant purposes and routines under the auspices of which the contents of the files are routinely assembled in the first place. Let the investigator attempt a remedy for shortcomings and he will quickly encounter interesting properties of these troubles. They are persistent, they are reproduced from one clinic’s files to the next, they are standard and occur with great uniformity as one compares reporting systems of different clinics, they are obstinate in resisting change, and above all, they have the flavor of inevitability. This inevitability is revealed by the fact that a serious attempt on the part of the investigator to remedy the state of affairs, convincingly demonstrates how intricately and sensitively reporting procedures are tied to other routinized and valued practices of the clinic. Reporting procedures, their results, and the uses of these results are integral features of the same social orders they describe. Attempts to pluck even single strands can set the whole instrument resonating.’ (Garfinkel, 1967, pp. 190 f.) That is, the ‘troubles’ arise whenever clinical records are used to ‘answer questions that depart in theoretical or practical import’ from the purposes for which they were assembled and external to the practices that for clinicians ‘are more or less taken for granted as right ways of doing things’. A clinic, like any enterprise, operates within a fixed budget and must, in its daily operation, consider the comparative costs of recording and obtaining alternative information. Some information items, such as sex and age of patients, are of course cheaply acquired, while other items, such as occupational history, require expensive reporting efforts (p. 192). At the same time, clinical records are assembled for future, variable, and generally unknown purposes. Consequently, such future purposes do not, in and of themselves, carry much weight in the busy daily life of the clinic. The division of labor in the clinic adds another source of ‘normal, natural troubles’: ‘The division of work that exists in every clinic does not consist only of differentiated technical skills. It consists as well of differential moral value attached to the possession and exercise of technical skills.’ For instance, the role records play in the accomplishment of administrative responsibilities is quite different from the role they play in the pursuit of professional medical responsibilities, and Garfinkel and Bittner pointed to ‘the wary truce that exists among the several occupational camps as far as mutual demands for proper record-keeping are con- Chapter 12: Frail foundations 405 cerned’ (p. 194). Thus, clinicians exhibit ‘abiding concerns for the strategic consequences of avoiding specifics in the record, given the unpredictable character of the occasions under which the record may be used as part of the ongoing system of supervision and review’ (p. 194). Now, the specific character of clinical work and the specific role of record keeping in these practices pose another source of trouble, a ‘critical source of trouble’ (p. 197). Clinical work consists in what Garfinkel and Bittner termed ‘remedial activities’. One of the crucial features of these is that ‘recipients [of treatment] are socially defined by themselves and the agencies as incompetent to negotiate for themselves the terms of their treatment’. Clinicians undertake to exercise that competence for them; they take responsibility for their patients. Accordingly ‘the records consist of procedures and consequences of clinical activities as a medico-legal enterprise’ (p. 198). This means that records are written and gathered for ‘entitled readers’. A ‘competent readership’ is presumed. The ‘contents of clinic folders are assembled with regard for the possibility that the relationship may have to be portrayed as having been in accord with expectations of sanctionable performances by clinicians and patients.’ (p. 199). Thus ‘terms, designations, and expressions contained in a document’ in the records are not ‘invoked in any “automatic” way to regulate the relationship’ of the terms to therapeutic activities. ‘Instead, the ways they relate to performances are matters for competent readership to interpret’ (p. 199). That is, ‘considerations of medicolegal responsibility exercise an overriding priority of relevance as prevailing structural interests whenever procedures for the maintenance of records and their eligible contents must be decided.’ So, although records may be put to uses that are different from those that serve the interests of considerations of medico-legal responsibility, ‘all alternatives are subordinated’ to considerations of medico-legal responsibility ‘as a matter of enforced structural priority’. ‘Because of this priority, alternative uses are consistently producing erratic and unreliable results’ (p. 200). All these conditions have important implications for the relationship between writer and reader of clinical records: ‘As expressions, the remarks that make up these documents have overwhelmingly the characteristic that their sense cannot be decided by a reader without his necessarily knowing or assuming something about a typical biography and typical purposes of the user of the expressions, about typical circumstances under which such remarks are written, about a typical previous course of transactions between the writers and the patient, or about a typical relationship of actual or potential interaction between the writers and the reader. Thus the folder contents much less than revealing an order of interaction, presuppose an understanding of that order for a correct reading.’ (p. 201). The records do work, however, because ‘there exists an entitled use of records’: ‘The entitlement is accorded, without question, to the person who reads them from the perspective of active medico-legal involvement in the case at hand and shades off from there. The entitlement refers to the fact that the full relevance of his position and involvement comes into play in justifying the expectancy that he has proper business with these expressions, that he will understand them, and will put them to good use. The specific understanding and use will be occasional to the 406 Cooperative Work and Coordinative Practices situation in which he finds himself. […] The possibility of understanding is based on a shared, practical, and entitled understanding of common tasks between writer and reader.’ (p. 201). That is, clinical records, especially records concerning legally touchy ‘remedial activities’, pose quite specific methodological challenges for secondary analytical use by non-competent readership. To emphasize this, Garfinkel and Bittner made a comparison with actuarial records: ‘A prototype of an actuarial record would be a record of installment payments. The record of installment payments describes the present state of the relationship and how it came about. A standardized terminology and a standardized set of grammatical rules govern not only possible contents, but govern as well the way a “record” of past transactions is to be assembled. Something like a standard reading is possible that enjoys considerable reliability among readers of the record. The interested reader does not have an edge over the merely instructed reader. That a reader is entitled to claim to have read the record correctly, i.e., a reader’s claim to competent readership, is decidable by him and others while disregarding particular characteristics of the reader, his transactions with the record, or his interests in reading it.’ (ibid., p. 202). Clinical records belong to practices quite different from actuarial records: ‘In contrast to actuarial records, folder documents are very little constrained in their present meanings by the procedures whereby they come to be assembled in the folder. Indeed, document meanings are disengaged from the actual procedures whereby documents were assembled, and in this respect the ways and results of competent readership of folder documents contrast, once more, with the ways and results of competent actuarial readership.’ (p. 203). The actuarial record ‘is governed by a principle of relevance with the use of which the reader can assess its completeness and adequacy at a glance.’ By contrast, with clinical records the reader, so to speak, reassembles the entries to ‘make the case’. It should be clear from this that no set of coding instructions, however elaborate and however meticulously designed, could have ameliorated the troubles Garfinkel and Bittner experienced. The troubles were found to be an inexorable feature of secondary use of clinical records. In other words, the trouble with following instructions in this case is, essentially, the kind of trouble one will expect to experience when engaged in reusing records that have produced for specific purposes within one work practice — outside of that practice and thus for purposes for which these records were not originally intended. Garfinkel and Bittner were indeed quite adamant to point this out. What they described is the kind of trouble historians engage in when they immerse themselves in the archives, the collections of internal memos, minutes, and private letters, etc. This kind of work is, essentially, interpretation work. These findings are important for sociological studies that depend on documentary evidence produced within a particular practice for local purposes, especially if issues of ethics and legal responsibility are at stake. They are also very informative with respect to the persistent troubles with ‘organizational memory’ and ‘knowledge management’ systems. But to construe Garfinkel’s argument as positing that instructions, by virtue of being linguistic constructs, always, everywhere, under all circumstances, are ‘incomplete’ and require ‘interpretative work’ and, Chapter 12: Frail foundations 407 hence, have ‘all of the ad hoc and uncertain properties that characterize every occasion of the situated use of language’, is preposterous. In spite of Garfinkel’s insistence that ethnomethodological studies are not meant to ‘encourage permissive discussions of theory’ (p. viii), Suchman’s interpretation turns an incisive analysis of certain methodological issues in investigating certain kinds of work practice into a philosophical proposition. That Suchman, at this crucial point in her argumentation, should have read Garfinkel in such a way is puzzling. Righteous fervor in the struggle against the cognitivist version of ‘rule governance’ would go some way towards explaining the urge to come up with a counter-theory. But there is also the special character of Garfinkel’s Studies in Ethnomethodology to consider. In this book Garfinkel focused on what one could call problematic situations, either ‘normally, naturally’ problematic ones like the coding case, or contrived ones like the breaching experiments and the counselling experiment. This focus is hand-in-glove with his objective: showing that the taken-for-granted assumptions of everyday life, the ‘natural attitude’, are researchable phenomena in their own right. This focus may, however, leave the impression that, according to Garfinkel, ordinary people, in the natural attitude of their daily work, struggle to make sense of coding schemes, production plans, administrative procedures, time tables. But that actors are not ‘judgmental dopes’ does not mean that they make it through the day by engaging in endless interpretive work and ad hoc activities.108 2.2. Sources of the interpretation myth in sociology Rule–skepticism is strong in the social sciences. There is an urge to interpose ‘interpretation work’ to account for practitioners’ acting in accordance with rules, plans, schemes, etc., — an urge so strong, in fact, that even Ryle’s and Wittgenstein’s demolition of this myth seems exceedingly difficult to grasp and accept. One source of this strong skepticist urge is the bafflement of a field worker when faced with myriad activities and inscriptions that do not seem to add up and make sense. It is natural for the field worker to project this bafflement and ascribe interpretation work to the observed practices: a natural fallacy. In the kinds of setting that are of primary concern to CSCW (cooperative work in organizational settings) the field worker will often find it difficult to align observable rule formulations (stipulated procedures, etc.) with the rules followed by practitioners. In such settings, the field worker will often find presumptive rule formulations that are not enforced, or seem to be in mutual contradiction, etc. One may furthermore come across rule formulations (stated procedures, work schedules, project plans) of which members seem ignorant or which they deliberately disregard. One may find rule formulations that seem to instruct actors to do a se108 It is only fair to point out that Suchman is not alone in this rule-skepticist generalization of Garfinkel. An article by Button and Harper offers a case in point (1995). The same abrupt generalization of Garfinkel’s very specific observations can be found in critical comment on Plans and Situated Actions by Sharrock and Button (2003). 408 Cooperative Work and Coordinative Practices ries of tasks in a specific order, for instance, ‘first do A, then B, and finally C’, but then members, sometimes or often, can be observed to jump from A to C while skipping B or to do B first and then A and C. Having experienced this, the field worker will be tempted to report that organizational rules do not instruct members what to do in a step-by-step manner, that they only convey general policies not operational guidelines, etc. An investigation of work practices that account for these practices in terms of the stated rules, by reference to the proverbial rule book, will obviously produce an utterly distorted account. In view of this, sociologists have adopted various strategies. Some will introduce a distinction between ‘formal’ and ‘informal’ organization or between ‘formal’ and ‘informal’ rules, etc. (e.g., Selznick, 1948). Others will effectively dissolve the very notion of rule-governed action by adopting the rule-skepticist position that work practices are ‘essentially ad hoc’ (e.g., Suchman, 1983, 1987; Bucciarelli, 1988a; Button and Harper, 1995). The field worker’s fallacy basically consists in mistaking the logical grammar of the concept of ‘rule formulations’ for that of ‘rules’ (Baker and Hacker, 2009, Chapter II). That we in our ordinary discourse do distinguish ‘rule formulations’ from ‘rules’ is evident. The same rule can be stated in different ways; it can be expressed orally, in writing, by gesturing, and so on; it can be formulated in different languages, by means of different notations, at different levels of detail, at different levels of formalization, by definition or by examples, by offering different examples, etc. One can make copies of rule formulations, but not of rules. However, the field worker’s rule–skepticist fallacy is also an natural one, in as much as we in our ordinary language do not always make a sharp distinction between the concept of ‘rule’ and ‘rule formulation’. In everyday life, when someone changes the formulation of a rule, the change will often be seen as a change of the rule. That is, in the words of Baker and Hacker, rules and rule-formulations cannot be simply segregated into ‘watertight compartments’, for ‘the grammars of “rule” and “rule-formulation” run, for a stretch, along the same tracks’ (Baker and Hacker, 2009, p. 47). As pointed out by Egon Bittner (1965), the investigator does not have privileged access to determining the governing sense of a stated rule (as formulated in a standard operating procedure or in a graphical representation of a classification scheme). The field worker is, by definition, an outsider to the setting; he or she does not (yet) understand the rule and has not (yet) been trained in applying the rule. In this the investigator is in the exact same situation as a novice being taught in the use of the same rule: like the novice, the investigator does not master the rule, hence the doubt, the uncertainty, the tentative applications. In other words, for lack of understanding, investigators are left with trying to interpret the stated rule. This is a practical-epistemological condition that makes field work serious work. It is important to keep in mind that there is nothing intrinsic in the form of a rule formulation that makes it a rule formulation; the sign does not need to have a Chapter 12: Frail foundations 409 specific form, such as, say, ‘If… then…’ or ‘Thou shall not…’. In the words of Baker and Hacker, again, ‘For the architectural historian or engineering students the blueprint describes the building or machine. It is used as a description; and if the building is other than as is drawn in the blueprint, then the blueprint is false. But for the builder or engineer the blueprint is used as an instruction or rule dictating how he should construct the building or machine. If what he makes deviates from the blueprint then (other things being equal) he has erred — built incorrectly’ (Baker and Hacker, 2009, p. 52). As field workers we cannot expect that rules are nicely stated in rule formulations that have the paradigmatic form of rule formulations; that is, we have to look at the actual practices of rule governed action and at the actual role of the various rule-stating artifacts (time tables, standard operating procedures, notation schemes, etc.) in those practices, to determine what the rule is. But this does not mean, of course, that ethnographic accounts by necessity are capricious or subjective. For a field worker trying to establish empirically what the rules of a particular practice actually are, the stated rule (the printed schedule, the procedure description, etc.) is of course an important source of data; but only one source among many, and a source that the field worker has no privileged access to understand. To establish what the rules in fact are, the field worker will have to consider how the stated rule is observably used in the setting. How are members instructed in applying the rule? How is it explained, exemplified, etc.? Furthermore, how is it invoked to justify or explain, justify, excuse, rectify, chastise, reprimand actions? How are actions that to an outsider might be seen an transgression of the stated rule actually treated by members? Are they approved or applauded; are they countenanced or condoned; or are they corrected, censured, castigated? In short, the task of determining the operative sense of a stated rule ‘is left to persons whose task it is to decide such matters’ (Zimmerman, 1966, p. 155). In all of this, the field worker is engaged in determining the rule of the practices empirically; their normative character to practitioners themselves can easily escape him or her. This is an insidious source of confusion and misrepresentation: the source of a natural fallacy. The fallacy arises when the inexorable investigational condition — that the operative sense of stated rules requires interpretation — is somehow construed as the human condition. That is, the interpretation work in which the field worker, as an outsider, is compelled to engage is conceived of as an inescapable condition for all, members and outsiders alike. The field worker’s fallacy consists in elevating his or her own mundane epistemological problems to the level of a human condition. This is of course confused and is a variant of the intellectualist legend: the conditions of intellectual work are the paradigm of the human condition and specifically intellectual practices the model of rational conduct. 410 Cooperative Work and Coordinative Practices 3. The consequences of counter-cognitivism Since CSCW aims at devising technologies that, when used, regulate aspects of interaction in cooperative work settings, and since work practices are both historically specific and specific to domains and settings, we need to understand the specificity of work practices: the rules, concerns, criteria, typifications, distinctions, priorities, notations, schemes, plans, procedures, etc.; how such rules etc. are applied unreflectively and unhesitatingly and also sometimes with some uncertainty; how they are sometimes questioned, debated, amended, elaborated; how they are taught and instructed; and how they evolve over time and how they propagate beyond local settings, are acquired, emulated, appropriated, etc.; what techniques practitioners bring to bear in their practices, how contingencies are dealt with routinely, how they employ the tools of the trade as intended in their design, how they often use them in ‘unanticipated’ ways, and how they often also seize incidental resources in the setting; and so on. For CSCW researchers to be able to do so, preconceived constructions of ‘human nature’ or ‘sociality’, applied as templates in technological design or in the production and analysis of ethnographic findings, would cause immediate sterility, since the specificity of cooperative work practices then would have been lost. It is thus of vital importance for CSCW to heed Egon Bittner’s call for ‘realism in field work’ and develop and maintain what he calls an ‘unbiased interest in things as they actually present themselves’ to practitioners (Bittner, 1973). For these reasons phenomenological sociology as represented by Alfred Schütz and the ethnomethodologists has played a very important role in CSCW. This was not, of course, preordained, as other intellectual traditions offer contributions that, for these purposes, are concordant with the phenomenological movement (e.g., the philosophies of Wittgenstein and Ryle, the tradition of ‘symbolic interactionism’, and in some respects also the psychology of Vygotsky). So far many if not most CSCW researchers will agree. The problem is that, to achieve this, ‘the field worker needs not only a good grasp of the perspectives of those he studies but also’, as pointed out by Bittner, ‘a good understanding of the distortive tendencies his own special perspective tends to introduce’. Because of the metaphysical tenet of Plans and Situated Actions, the version of ethnomethodology that has been widely received in CSCW is one in which phenomenological sociology has been deprived of its most important insights: the principle of specificity of practices (‘zones of relevance’, ‘finite provinces of meaning’, etc.), the notion of the ‘natural attitude’ of working, the principle of taking the point of view of practitioners (what are they up to? how does the world look when on does that kind of work?), the method of conceiving of work practices in practitioners’ own terms, as opposed to through universal constructs, externally imposed criteria, etc. While Suchman’s book showed a way out of cognitivism, towards studies of actual work practices, it also — unwittingly but effectively — was instrumental in establishing another dogma, the dogma that plans play no role in action, i.e., in Chapter 12: Frail foundations 411 determining the course action. The dogma was not intended but was the consequence of Suchman’s mirroring cognitivism’s metaphysical form of reasoning. The existence of this dogma is evident. Consider, for example, this admonishment from a widely cited article written by Paul Dourish and Graham Button: ‘The disturbingly common caricature of her position is that there are no plans, but only “situated actions” — improvised behaviors arising in response to the immediate circumstances in which actors find themselves and in which action is situated. In fact, as Suchman has been at pains to point out, she did, in fact, accord an important status to plans as resources for the conduct of work. Her argument was that plans are one of a range of resources that guide the moment-by-moment sequential organization of activity; they do not lay out a sequence of work that is then blindly interpreted.’ (Dourish and Button, 1998, pp. 405 f.) This defense of Suchman stayed within the metaphysical discourse established by Suchman’s original argument: ‘plans […] do not lay out a sequence of work that is then blindly interpreted’. Disregarding the perplexing expression ‘blindly interpreted’ (what would it mean to ‘interpret blindly’?), I take it that the authors meant to say ‘plans do not lay out a sequence of work that is then blindly [followed]’. Well, plans sometimes are. In fact, they are most of the time. They are routinely applied as unproblematic guidelines or instructions, and if plans do not lay out a sequence of work that is then normally followed without reflection then there are no plans as we ordinarily understand the term. The statement that ‘plans do not lay out a sequence of work that is then blindly [followed]’ only makes sense if ‘plans’ are not really ordinary plans, but rather ‘plans’ as conceived of in cognitivist theorizing. But then ‘plans’ as conceived of in cognitivism cannot be followed (or ‘interpreted’), for they are obscure causal mechanisms. In another widely cited article Button and Harper cautioned ‘designers’ in CSCW not to misunderstand the concept of ‘work practice’ (Button and Harper, 1995). Their concern was occasioned by the increased use of the concept of ‘work practice’ in the ‘design community’, especially in the wake of Suchman’s Plans and Situated Action. They noted with barely suppressed irritation that the concept of ‘work practice’ was ‘being invoked more and more in the rhetoric that surrounds design’ (p. 266) and had become ‘something of a rallying cry in many quarters of CSCW’ (p. 279); but what gave them particular cause for concern was that the concept of ‘work practice’ had its ‘origins in sociology’ and had ‘a well established place in the sociology of work’ and that the ‘sociological underpinnings’ of the concept and ‘the order of work organisation to which Suchman refers’ might not be properly understood (p. 263-265). The problem, according to Button and Harper, was that the concept of ‘work practice’ in sociology of work was used for ‘describing amongst other things the rule book formulations of work as well as the situated responses to contingent circumstances’ (p. 265, emphasis added). The authors’ issue with this use of the concept of ‘work practice’ seems to be that ‘rule book formulations of work’ are even considered in accounts of work practices. They contrasted this with what they presented as the ethnomethodological position. Referring to Garfinkel and Bittner’s study of the ‘normal, natural troubles’ of coding clinical records they made the following claims: 412 Cooperative Work and Coordinative Practices ‘The finding of Garfinkel’s study was […] that in practice it was not possible to exhaustively and explicitly stipulate the coding rules. However full and detailed the rules in the coders’ handbook were made, each time coders had to administer the schedule, there was a need for decision and discretion. The coders would resort to a variety of practices to decide what the coding rules actually required of them and whether what they were doing was actually (or virtually) in correspondence with those rules. These were essentially ad hoc relative to the coding manual’s purportedly systematic character. It was through the implementation of these ad hoc practices that coders achieved their work of coding. The formalised account of the work of coding as applying the rules omits the very practices that organise that work.’ (Button and Harper, 1995, p. 265). According to Button and Harper, then, ‘work-practices constitute, in its [sic] detail and in the face of the unfolding contingencies of work, the temporal order of work and the ordinary facticity of domains of work’ (p. 264). In short, practices are ‘essentially ad hoc’. Now, it should be said immediately and with emphasis that this pallid notion of practice has not prevented these authors from investigating complex professional practices and, in doing so, blessing CSCW with some of the most influential studies of work practices (e.g. Harper, et al., 1989a, b; Bowers, et al., 1995). Harper and Button have provided us with very insightful analyses of normatively regulated conduct in ordinary work settings. What we have, then, is what looks like a clean separation of ‘theory’ and ‘practice’, the hallmark of a dogma. Its paralyzing effects show up elsewhere. The upshot of all this is that Plans and Situated Actions has been instrumental in replacing the cognitivist dogma with another. The new dogma, while certainly far more fertile in terms of encouraging empirical work than that of the cognitivist wasteland, made CSCW largely numb to the massive web of coordinative practices and techniques that characterizes typical modern workplaces. Their presence is, of course, not flatly denied; it is even acknowledged that they serve as ‘important’ ‘resources’ for situated action. But it is, ab initio and dogmatically, posited that ‘plans do not lay out a sequence of work that is then blindly [followed]’, and that work practices are ‘essentially ad hoc’. The concept of ‘practice’, which plays a defining role for CSCW’s research program, is hereby emptied of content, rendered useless. The practice-oriented research program of CSCW is effectively undermined. The implication is a strong methodological bias even in ethnographic fieldwork. To see this, take for instance the Bob Anderson’s support of Suchman against her critics: ‘If we set the context for the ethnography at the level at which many ethnographers feel most comfortable, we will find they are almost obsessed with change of one sort or another. In picking their way through the minutiae of routine action, prominence is (endlessly) given to the innovative, the ad hoc, and the unpredictable rife in the workplace and elsewhere. Change, here, is the very stuff of ethnography.’ (Anderson, 1997, p. 177, emphases added). This is an accurate statement of the diagnosis. However, Anderson does not mean this as a critique but as a formulation of an optional methodological preference. Chapter 12: Frail foundations 413 What Anderson does not take into account is Egon Bittner’s warning, in the early days of ethnomethodology, against the fieldworker’s fallacy: Because the ethnographer is ‘a visitor whose main interest in things is to see them’, to him or her ‘all things are primarily exhibits’; they are ‘never naturally themselves but only specimens of themselves’ (Bittner, 1973, p. 121). These are unavoidable conditions for any ethnography. However, as a result, a certain intellectualism may distort the ethnography: ‘Since the field worker […] always sees things from a freely chosen vantage point — chosen, to be sure, from among actually taken vantage points — he tends to experience reality as being of subjective origin to a far greater extent than is typical in the natural attitude. Slipping in and out of points of view, he cannot avoid appreciating meanings of objects as more or less freely conjured.’ (Bittner, 1973, pp. 121 f.). And in doing so, the field workers describes the setting ‘in ways that far from being realistic are actually heavily intellectualized constructions that partake more of the character of theoretical formulation than of realistic description’ (Bittner, 1973, pp. 123 f.). Bittner is here restating and underscoring Winch’s fundamental proposition, namely, ‘it is not open’ to the sociological investigator ‘to impose his own standards from without. In so far as he does so, the events he is studying lose altogether their character as social events’ (Winch, 1958, p. 108). If we in our analyses of work practices give prominence to ‘the innovative, to the ad hoc, to the unpredictable rife in the workplace’, or to the ‘unfolding contingencies of work’ we will be under the ‘serious misunderstanding’ of rendering practices in ways that are ‘heavily intellectualized constructions.’ That is, if we heed Bittner’s warning, we cannot take Anderson’s rather agnostic position and simply accept the reputed ‘obsession with change’ as a methodological preference some may wish to adopt and others not. For an ethnography characterized by an ‘obsession with change’ is a ‘heavily intellectualized construction’ just as much as one obsessed with stable structures, equilibrium, and what not. Obviously, such ‘heavily intellectualized constructions that partake more of the character of theoretical formulation than of realistic description’ can only impede CSCW’s ability to address the construction and use of plans in cooperative work, for the ordinary plans of ordinary cooperative work in ordinary organizational life are then rendered delusory, the natural attitude of practitioners excluded, if they are seen at all. In sum, the consequence of all this is, at best, that our studies become unrealistic. At worst, CSCW abandons its practice-oriented technological research program and becomes irrelevant as yet another research area producing post hoc descriptions of the various uses of new products, services, and facilities coming onto the market. 4. The problem of computational artifacts I have focused on Suchman’s account for obvious reasons. Her contribution is, justifiably, seen as a major contribution to the intellectual foundation of CSCW 414 Cooperative Work and Coordinative Practices and HCI. By centering on the relationship between the concepts of ‘plans’ and ‘situated action’, it provided the foundational conceptual apparatus of CSCW: its axis of conceptualization. However, this conceptualization is no longer holder up. Confusion reigns with respect what constitutes CSCW’s problem and its scope. In fact, CSCW’s central problem, its axis of conceptualization, can not be adequately defined in terms of ‘plans and situated action’. CSCW’s program is not usefully conceived of in terms of ‘plans and situated action’ or similar, for the problem CSCW has to address is not primarily how normative constructs such as ordinary plans and other organizational rules are applied in practice but how practitioners use or may use machines in following the rules of their cooperative work practices (schedules, procedures, categorizations, schemes). And that is an entirely different issue. I will suggest, therefore, that CSCW’s problem is circumscribed by the concepts of rule-governed work practices on one hand and on the other highly regularized causal processes in the form of ‘computational artifacts’ used, within cooperative work practices, for purposes of coordinating interdependent activities. How do we incorporate mechanical regulation of action and interaction in cooperative work practices? How can we exploit the immense flexibility of computational technology to give ordinary workers effective control over the distributed execution, construction, and maintenance of mechanized coordinative protocols? The point of Suchman’s critique of cognitivism was of course to be able to address the problem of computational artifacts. The problem she set out to address was not the role of plans in situated action in general but the problem of how plans incorporated in computational artifacts (‘reified in the design of intelligent machines’) affect human action and interaction, how they can be appropriated in our practices, and what implications for design may be derived from that: ‘I have attempted to begin constructing a descriptive foundation for the analysis of human-machine communication’ (p.180). Based on her study of a problematic case of human-computer interaction (a computer-based system intended to instruct users of a complex photocopier), Suchman concluded: ‘The application of insights gained through research on face-to-face human interaction, in particular conversation analysis, to the study of human-computer interaction promises to be a productive research path. The initial observation is that interaction between people and machines requires essentially the same interpretive work that characterizes interaction between people, but with fundamentally different resources available to the participants. In particular, people make use of a rich array of linguistic, nonverbal, and inferential resources in finding the intelligibility of actions and events, in making their own actions sensible, and in managing the troubles in understanding that inevitably arise. Today’s machines, in contrast, rely on a fixed array of sensory inputs, mapped to a predetermined set of internal states and responses. The result is an asymmetry that substantially limits the scope of interaction between people and machines. Taken seriously, this asymmetry poses three outstanding problems for the design of interactive machines. First, the problem of how to lessen the asymmetry by extending the access of the machine to the actions and circumstances of the user. Secondly, the problem of how to make clear to the user the limits on the machine’s access to those basic interactional resources. And finally, the problem of how to find ways of Chapter 12: Frail foundations 415 compensating for the machine’s lack of access to the user’s situation with computationally available alternatives.’ (Suchman, 1987, pp. 180 f.). Suchman summarized the conclusion as follows: ‘I have argued that there is a profound and persisting asymmetry in interaction between people and machines, due to a disparity in their relative access to the moment-by-moment contingencies that constitute the conditions of situated interaction. Because of the asymmetry of user and machine, interface design is less a project of simulating human communication than of engineering alternatives to interaction’s situated properties’ (Suchman, 1987, p. 185). The metaphor of ‘asymmetry’ is perplexing. Would it make sense to say that my relationship to my alarm clock is ‘asymmetrical’ because it seems completely insensitive to the effects of last night excesses, due to ‘a disparity’ in our ‘relative access to the moment-by-moment contingencies’? Does a notion of ‘asymmetry’ not presume a common metric, some significant commonality? In which way, then, are the two parties commensurate? I take it that Suchman would have answered those expressions of disbelief by emphasizing what she already said in the quote above: that ‘interaction between people and machines requires essentially the same interpretive work that characterizes interaction between people, but with fundamentally different resources available to the participants’. This is truly bewildering: ‘essentially the same interpretive work’! Now, Suchman certainly did not intend to advocate a cognitivist position but rather to provide an ‘alternative account’, but something was fatally amiss in this line of reasoning. The problem, or so it seems to me, is that she conceived of computational artifacts as linguistic constructs: ‘I argue that the description of computational artifacts as interactive is supported by their reactive, linguistic, and internally opaque properties’ (pp. 7, 16). On that view, it was not a dramatic jump to assert that ‘interaction between people and machines requires essentially the same interpretive work that characterizes interaction between people’. But then the alternative account to cognitivism has already been effectively abandoned. The containment strategy turns out not to contain cognitivism at all. The high ground has been evacuated, the major highway intersections long since lost, defeat all but conceded. The strategy has landed us in something akin to a Green Zone, surviving at the mercy of the adversary we proudly set out to vanquish. Still, the notion that computational artifacts have ‘linguistic properties’ is, of course, one that is taken for granted in the ordinary discourse of computer science where it is in common usage (witness terms like ‘programming language’ etc.). There, in the context of everyday reasoning about programming, it is (largely) unproblematic, as harmless and pragmatic as vernacular expressions such as ‘sunrise’ and ‘sunset’. However, outside of the discourse of computer science, and especially in the context of reasoning about human-computer interaction and computer-supported cooperative work, the notion of computational artifacts as linguistic constructs is fatally misleading. This should be evident inasmuch as Suchman conceived of computational artifacts in exactly the same terms as she conceived of ‘plans’: as ‘essentially linguistic representations’ (p. 186). It is a category mistake of the first order. 416 Cooperative Work and Coordinative Practices The problem is, as should be clear by now, that the seeds of the defeat were there from the very beginning, inherent in Suchman’s strategy of trying to demolish cognitivism while accepting cognitivism’s metaphysical discourse and its mechanistic premises. In the cognitivist account, there is no room, i.e., no logical space, for considering normative regularity. There is only room for empirical observations of concomitance (and a significant amount of speculative theorizing, of course). On this view, ‘rules’ can only be conceived of as hypothetical propositions, derived inductively, by correlation, from observed regularity. The notion that rules, plans, etc., are an essential part of our practices, as public standards of correctness or incorrectness, is unintelligible on this view. Ordinary normative activities, such as asking for the way to the nearest grocery shop, suggesting a definition of a word, setting the table for a dinner party, executing a production plan, etc. have no place in this discourse. Suchman’s ‘alternative account’ unfortunately left this view unchallenged. As a result, the normative nature of our practices was lost in the fire. For the same reason, computational artifacts are ascribed properties of the normative category, which means that we end up with trying to understand ‘human-computer interaction’ with the key concepts of ‘practices’ and ‘computational artifacts’ completely confounded. 417 Chapter 13 Dispelling the mythology of computational artifacts ‘If calculating looks to us like the action of a machine, it is the human being doing the calculation that is the machine.’ (Wittgenstein, 1937-44, IV § 20) ‘Turing’s “Machines”. These machines are humans who calculate.’ (Wittgenstein, 1946-49, § 1096) What is paralyzing CSCW is the assimilation of the normative concept of plans, schemes, schedules, and similar organizational constructs with the mechanist concept of causal determination of rational action. This leaves no conceptual room for CSCW. Given this confusion, any thought anybody may have of building or studying computational artifacts that embody plans, schemes, schedules, will unavoidably be met with disbelief or disinterest: ‘But we know already that that’s wrong/impossible, for action is essentially ad hoc!’ So, instead the good CSCW researcher will focus on building or studying computational artifacts that embody as little computational regulation of human interaction as possible. The ideal, one might say, is computational artifacts on the model of plasticine. The root of these problems is the concept of computational artifacts, or rather: the philosophy of computing, a channel of endless mystification. To take just one example, from a book by a well-respected computer scientist: ‘I etch a pattern of geometric shapes onto a stone. To the uninitiated, the shapes look mysterious and complex, but I know that when arranged correctly they will give the stone a special power, enabling it to respond to incantations in a language no human has ever spoken. I will ask the stone questions in this language, and it will answer by showing me a vision: a world created by my spell, a world imagined within the pattern of the stone.’ (Hillis, 1998, p. vii) The interesting thing is that Hillis then goes on to give an instructive account, completely free from such, well, incantations, of the concept of computing, showing how computing operations can be performed by means of causal processes in mundane (mechanical, hydraulic) devices. 418 Cooperative Work and Coordinative Practices What mystifies us is the ‘mythology of a symbolism’, as Wittgenstein puts it: in this case the mythology of computational artifacts. It is, prima facie, a mythology of our everyday dealings with computational artifacts. What mystifies us is, first of all, the computer’s apparent vivacity: the impression of an infinite number of possible internal states and its dynamic reactivity. It is an unavoidable aspect of using computers that the user engages in forms of thinking in which ‘interacting’ with electronic processes is considered natural; it is, in a strict Schutzian sense, the user’s natural attitude. As computer users and technicians we engage — as the most natural thing in the world — in ascribing linguistic and other anthropomorphic properties to computational artifacts: programming ‘language’, ‘memory’, ‘information’, ‘ontology’. The computational artifact mystifies us in the same way as clocks and their ability to ‘tell time’ for ages have allured thinkers to speculate about the mystery of time. This mystification is unavoidable even if one knows that the time piece executes a causal process in a manner that is carefully devised to be exceptionally regular, and that it is this regular causal process that allows us to use the clock as a standard metric for measuring the sequential order or irreversible nature of other casual processes, from heart beats to train movements (Gell, 1992). A clock does not ‘tell’ us anything more about time than the meter rod ‘tells’ us about space. But the meter rod in Paris, which has served as standard referent for measuring spatial properties of things, does not mystify us, because it does not do anything. The clock, however, seems to do something when its springs uncoil, its gears turn, and its hands move; it is an automaton, and we therefore unhesitatingly ascribe the property of ‘telling time’ to it. Or to take another, more mundane, example, there are devices for sorting eggs according to their size. The eggs will roll along a slightly declining plane that has holes in different, but increasing sizes: small eggs will drop into the first holes, medium size eggs into the next ones, and so on, until only XL eggs remain. The device does not measure the size of eggs; the operator measures the size of eggs by means of the device, according to the standards agreed to by the egg industry. What makes us engage in ‘incantations’ and ascribe ‘egg-sorting’ behavior to the ‘egg sorter’ is that the ‘egg sorter’ is, within the horizon of sorting eggs, selfacting. It is the automatic performance — the allocation of the control function to the technical implement — that prompts us to engage in such ascription of function. Another source of mystification is the manner in which the software machine is built: it is built by a process that, to the programmer and his or her admiring observers, is a linguistic activity. The programmer writes text: code. What happens to the code afterwards is hidden from view, out of mind, for it is executed automatically: the transformation of the ‘source code’ by the compiler into ‘object code’ in binary format. In this process, a linguistic construct, the source code, is transformed into a binary machine that, in the form of electronic patterns, can loaded into RAM and then executed in conjunction with the hard-wired machinery of the CPU. The source code is not a machine but the blueprint for one; but in contrast to an ordinary blueprint, which is also a linguistic construct, the source Chapter 13: The mythology of computational artifacts 419 code is not handed over to a machinist who then builds the intended machine but is submitted to a special software machine, the compiler, that automatically performs the construction of the intended machine as specified in the blueprint or source code.109 A deeper source of the mystification surrounding the computer, or rather: a more insidious source, is that we routinely use software machinery, i.e., regimented causal processes, as an integrated aspect of our normative or rule-following practices and that we are tempted to assimilate the two — the causal process and our use of it — and become mystified, struck with awe at the fruit of our hands. Just as it is not clocks that ‘know’ the time or ‘tell us’ what time it is, but we who use clocks to ‘tell the time’, it is not the computer that computes, calculates, executes plans, searches for information, etc., but we who apply computational artifacts to do so in our normative practices of computing, calculating, executing plans, searching for information, etc., just as we have previously employed abacuses, slide rulers, desktop calculators, time pieces, filing cabinets, etc. 1. Computational artifacts, or Wittgenstein vs. Turing The source of this conceptual muddle is, again, the failure to grasp the categorial distinction between normative behavior and mere regularity. To sort out this mess, which in the case of computational artifacts is exceptionally dense, we are in the fortunate situation that Wittgenstein was developing a critical conceptual analysis of the concept of calculation or computation that was then, in the late 1930s, being developed by Alan Turing and others. It is fairly well-known that Turing and Wittgenstein knew each other from 1937, that Turing followed Wittgenstein’s lectures on the foundations of mathematics (Wittgenstein, 1939), and that they in the course of these lectures engaged in discussions concerning the nature of mathematics. What is less well known is that Turing sent a reprint of his paper to Wittgenstein in February 1937 (Copeland, 2004, p. 130) and that Wittgenstein, in a bulky collection of manuscripts on the philosophy of mathematics written from 1937 to 1944, subjected (inter alia) Turing’s mechanistic framing of his argument to incisive critical discussion (Wittgenstein, 1937-44).110 A few remarks on the background and general context are required, however. Turing’s machine was devised in the context of the ‘foundations crisis’ that mathematicians and philosophers then believed had struck mathematics, and Wittgenstein’s critique of Turing and his mechanist thesis was but a theme in his critical 109 This linguistic mystique also appears when a user ‘interacts’ with the computational artifact by writing ‘commands’ that then activate an entire system of machines under the control of the computer’s operating system. It is worth noticing that such a mystique is absent when a user pushes the start button on a dish washer. 110 For very useful commentaries on Wittgenstein’s critique of Turing in particular and the ‘mechanist thesis’ in general, cf. the penetrating studies by Stuart Shanker (1987c, d, 1995, 1998). 420 Cooperative Work and Coordinative Practices discussion of the very notion that such as crisis existed and of the attempts to overcome it. 1.1. Foundations lost The conceptual underpinnings of computing technologies (symbolic logic, algorithmics, recursive functions, etc.) grew out of techniques that were developed in the last decade of the 19th and the first decades of the 20 centuries in an attempt to establish mathematics on a complete and consistent foundation. By the end of the 19th century, leading mathematicians had become deeply concerned with the foundations of their science. It was felt by many that the impeccable certainty and consistency of mathematical knowledge could be questioned. It was widely perceived as a ‘foundations crisis’. Developments in mathematics such as non-Euclidian geometry (by Lobachevski, Riemann, and others) and Cantor’s theory of sets and ‘transfinite numbers’ had made it problematic for many mathematicians to continue to rely on intuition in proving theorems. Their concerns became acute when paradoxes and absurdities began to emerge. To save mathematics some mathematicians (especially L. E. J. Brouwer) went so far as to state that only those mathematical proofs that could be constructed by ‘finite’ methods, i.e., by definite and surveyable steps, and which therefore were deemed ‘intuitively’ evident, were acceptable as bona fide members of the body of mathematics. Going under the name ‘intuitionism’ or ‘constructivism’, this program was prepared to jettison whatever could not be proved by finite methods, although this would chuck off large chunks of highly useful mathematics. Unsurprisingly this Procrustean program was deemed unattractive by most of the mathematicians that participated in trying to overcome the ‘foundations crisis’. The main axis of attack on the crisis was developed by the British philosopher Bertrand Russell. He saw mathematics as an ‘edifice of truths’, ‘unshakable and inexpugnable’, a vehicle of reason that lifts us ‘from what is human, into the realm of absolute necessity, to which not only the actual world, but every possible world, must conform’ (Russell, 1902, pp. 69, 71). Little wonder that he felt an urge to safeguard mathematics. In his effort to do so, he developed a research program that in the foundational debates became known as ‘logicism’, a program to rebuild mathematics into a monolithic structure of propositions based on a drastically limited set of axioms and rules of inference. Building on the work of Frege and Peano, he first made an impressive attempt to reconstruct mathematics on the basis of first-order logic in the form of a theory of ‘classes’ or sets (Russell, 1903). However, he had barely finished the book before he realized that the edifice he had erected to salvage the ‘realm of absolute necessity’ by establishing mathematics on the unshakeable basis of logic — was deeply flawed: the concept of ‘class’ derived from Cantor’s set theory led to absurdities. Mathematicians like David Hilbert were dismayed: ‘In their joy over the new and rich results, mathematicians apparently had not examined critically enough whether the modes of inference employed were admissible; for, purely through the ways in Chapter 13: The mythology of computational artifacts 421 which notions were formed and modes of inference used — ways that in time had become customary — contradictions appeared, sporadically at first, then ever more severely and ominously. They were the paradoxes of set theory, as they are called. In particular, a contradiction discovered by Zermelo and Russell had, when it became known, a downright catastrophic effect in the world of mathematics.’ (Hilbert, 1925, p. 375) Russell spent most of a decade — together with Alfred Whitehead — building another foundation of mathematics (Whitehead and Russell, 1910). In order to exclude paradoxes, the new foundation was built by means of an elaborated logical calculus called the ‘theory of types’, ‘a complicated structure which one could hardly identify with logic’ in the normal sense (Davis and Hersh, 1980, p. 333). In Hilbert’s words, ‘Too many remedies were recommended for the paradoxes; the methods of clarification were too checkered’ (Hilbert, 1925, p. 375). In an effort to avoid these problems Hilbert developed another approach, generally known as ‘formalism’. Where Russell dreamt of restoring the ‘realm of absolute necessity’, Hilbert’s program was predicated on the ‘conviction’ that every mathematical problem is solvable: ‘We hear within us the perpetual call: There is the problem. Seek its solution. You can find it by pure reason, for in mathematics there is no ignorabimus’ (Hilbert, 1900, p. 248). For Hilbert, therefore, the situation was intolerable: ‘Let us admit that the situation in which we presently find ourselves with respect to the paradoxes is in the long run intolerable. Just think: in mathematics, this paragon of reliability and truth, the very notions and inferences, as everyone learns, teaches, and uses them, lead to absurdities. And where else would reliability and truth be found if even mathematical thinking fails?’ (Hilbert, 1925, p. 375). To salvage mathematics as an unshakeable edifice, as the ‘paragon of reliability and truth’, Hilbert formulated a grandiose ‘formalist’ program to demonstrate or ensure — in the most rigorous manner — that the corpus of mathematical calculi constitute a complete, consistent, and decidable formal system. His vision was to develop mathematics ‘in a certain sense’ ‘into a tribunal of arbitration, a supreme court that will decide questions of principle and on such a concrete basis that universal agreement must be attainable and all assertions can be verified’ (Hilbert, 1925, p. 384). Hilbert’s strategy was to regain the certainty of ‘intuitive’ perceptual inspection even in regions of mathematics such as ‘transfinite number’ theory that defied perceptual inspection, and to do so by treating ‘formulas’ as ‘concrete objects that in their turn are considered by our perceptual intuition’ (Hilbert, 1925, p. 381). In other words, the approach he proposed was to encode mathematical propositions in a ‘formal’ language, that is, (a) to ‘replace’ ‘contentual references’ by ‘manipulation of signs according to rules’, (b) treat ‘the signs and operation symbols as detached from their contentual meaning’ and (c) thereby ultimately obtain ‘an inventory of formulas that are formed by mathematical and logical signs and follow each other according to definite rules’ (Hilbert, 1925, p. 381). Now, Hilbert most certainly did not think of mathematics as a meaningless game (in epistemological matters, he was a ‘realist’ or Platonist); but, to stave off impending skepticism, he was ready to advocate this radically formalistic approach 422 Cooperative Work and Coordinative Practices as a methodology, and as a result his formalist program was accused of treating mathematics as a meaningless game. Unruffled, Hilbert retorted by appropriating the epithet: ‘What, now, is the real state of affairs with respect to the reproach that mathematics would degenerate into a game? […] This formula game enables us to express the entire thought-content of the science of mathematics in a uniform manner and develop it in such a way that, at the same time, the interconnections between the individual propositions and facts become clear’ (Hilbert, 1927, p. 475). Although Hilbert’s program took a different course than Russell’s, the two programs shared basic premises, namely that the problems facing mathematics at the time were epistemological problems: how could mathematicians ensure the impeccable and unassailable validity of the propositions of higher mathematics where ‘perceptual intuition’ seemingly had lost traction (‘transfinite numbers’ etc.)? What united the two camps was the belief that mathematics was in need of epistemological underpinnings and that recent extensions to mathematics made the received epistemological foundations questionable, as they no longer ensured ‘self-evident’ truths. And both programs shared the belief that if it could be shown that the whole of mathematics could be derived from a few ‘self-evident’ axioms and rules of inference, in a finite sequence of steps, then the state of impeccable certitude could be regained. Wittgenstein played a key role in the development of mathematical logic and in the development of the philosophy of mathematics,111 but did not share the sense of crisis. There was, in his view, nothing in the development of mathematics that should cause angst among mathematicians. In the midst of the intellectual turmoil he commented coolly: ‘I have the impression that the whole question has been put wrongly. I would like to ask, Is it even possible that mathematics to be inconsistent?’ (Wittgenstein, 1929-32, p. 119). Observing the acute fear of deep-seated inconsistencies and antinomies, he commented that a contradiction ‘is only a contradiction if it is there’: ‘People have the notion that a contradiction that nobody has seen might be hidden in the axioms from the beginning, like tuberculosis. You do not have the faintest idea, and then some day or other the hidden contradiction might break out and then disaster would be upon us’ (ibid., p. 120). Russell’s antinomies (e.g., his infamous paradox) had of course caused anxiety and wrought havoc, but in Wittgenstein’s view these antinomies ‘have nothing whatsoever to do with the consistency of mathematics; there is no connection here at all. For the antinomies did not arise in the calculus but in our ordinary language, precisely because we use words ambiguously’ (ibid., p. 121). On Wittgenstein’s diagnosis, the problem was not a problem within mathematics but was rooted in deep confu111 In fact, the philosophy of mathematics was Wittgenstein’s central concern from the Tractatus until about 1945 (when he switched his attention to another muddle: the philosophy of psychology). This is evidenced by the records of his discussions with members of the Vienna Circle (1929-32), the manuscripts published as Philosophical Remarks (1930) and Philosophical Grammar (1931-34), his lectures at Cambridge (1932-33, 1939), his Remarks on the Philosophy of Mathematics (1937-44), and of course the Philosophical Investigations (1945-46). Indeed, it was his intention that the second volume of Philosophical Investigations should be based on the later manuscripts on mathematics. Chapter 13: The mythology of computational artifacts 423 sion in the way mathematicians and philosophers alike interpreted mathematics: ‘If I am unclear about the nature of mathematics, no proof can help me. And if I am clear about the nature of mathematics, then the question about its consistency cannot arise at all’ (ibid.). The source of the confusion, Wittgenstein found, was the view, shared by both logicists and formalists, that mathematics is an investigation that, on par with physics, investigates a mathematical reality that exists independently of mathematical practice and which serves to adjudicate the correctness of that practice and language. A major part of his critical examination of the philosophy of mathematics was therefore devoted to exposing these dearly held epistemological assumptions (Shanker, 1986b, 1987a; Gerrard, 1991). The overriding aim of his argument with philosophy of mathematics was to show that mathematical propositions are not empirical propositions about preexisting objects but, rather, that ‘mathematics forms a network of norms’ (Wittgenstein, 1937-44, VII §67). Mathematical propositions do not appear ‘hard’, ‘inexorable’, nonnegotiable, because they are crystallizations of experiential evidence, for then they would be contingent; they appear ‘hard’ because they are normative: If one says ‘If you follow the rule, it must be like this’, then one does not have ‘any clear concept of what experience would correspond to the opposite’, for ‘the word “must” surely expresses our inability to depart from this concept’. The statement that ‘it must be like this’, does not mean, ‘it will be like this’: ‘On the contrary: “it will be like this” chooses between one possibility and another. “It must be like this” sees only one possibility’. Consequently, ‘By accepting a proposition as self evident, we also release it from all responsibility in face of experience.’ This, Wittgenstein argued, is the source of the ‘hardness’ of mathematical propositions (IV, §§ 29-31). Mathematical propositions are rules, they specify the grammar of number words. Still, mathematical propositions are rules of a special kind, namely rules whose certainty is established by ‘proofs’. There is nothing otherworldly about the proof and its objectivity: the proof is not — like a physical law — based on experiential evidence that can be produced to justify it; is we who accept the rule and accept to be guided the rule. The mathematical proposition can be said to be true in much the same way that one can ask: ‘Is it true that there are 29 days in February this year?’. That is, its truth is grammatical, it refers to the actuality of a conventionally constituted fact. But the rule expressed by a mathematical proposition is not the result of an arbitrary choice; the proof ‘confers certainty’ upon the proposition by ‘incorporating it — by a network of grammatical propositions — into the body of mathematics’ and thereby making it incontestable (Hacker, 1989, p. 333). The proof establishes a connection between mathematical concepts and provides a concept of the connections. The proof ‘introduces a new concept’; ‘the proof changes the grammar of our language, changes our concepts. It makes new connexions, and it creates the concept of these connexions. (It does not establish that they are there; they do not exist until it makes them.)’ (Wittgenstein, 1937-44, III §31). 424 Cooperative Work and Coordinative Practices In accordance with this view of mathematics — as a network of concepts — Wittgenstein emphasized that mathematical calculi could not be reduced to one formalism: ‘After all, the natural numbers are not identical with the positive integers, as though one could speak of plus two soldiers in the same way that one speaks of two soldiers; no, we are here confronted with something new’ (Wittgenstein, 1929-32, p. 36). That is, the concept of the natural number 2, the 2 concept of the integer +2, and the rational number 1 are not the same concepts; their grammars are distinctly different (cf. Waismann, 1930; Waismann, 1936, pp. 60 f.). One can subtract two people from the list of guests to be invited for dinner but one cannot have a negative number of€guests for dinner. So, pointing to the grammatical heterogeneity of mathematics, Wittgenstein objected vehemently against the very idea that mathematics be subjected to forced formalization: ‘“mathematics” is not a sharply delimited concept’; rather, ‘mathematics is a motley of techniques and proofs’ (Wittgenstein, 1937-44, III §§46, 48), and the ‘invasion of mathematics by mathematical logic’ is a ‘curse’, ‘harmful’, a ‘disaster’ (Wittgenstein, 1937-44, V §§ 24, 46). Consequently, Wittgenstein saw ‘mathematical logic [as] simply part of mathematics. Russell’s calculus is not fundamental; it is just another calculus’ (Wittgenstein, 1932-33, p. 205). He similarly argued that what Hilbert thought of a ‘metamathematics’ was just another mathematical calculus next to the others and that the attempt to reduce the ‘motley’ of mathematics to that single calculus was harmful: ‘There is no metamathematics’ (Wittgenstein, 1931-34, p. 296). The point of Wittgenstein’s critique of mathematical philosophy was not to denigrate or invalidate any part of mathematics. On the contrary. It deliberately and carefully left mathematics as it was, only — perhaps — unburdened of some of the metaphysical excretions that had accumulated in the form of mathematicians’ prose interpretations of their achievements and problems and in the form of philosophical ‘invasions’. The epistemological angst that gripped mathematicians and philosophers of mathematics a century ago has long since dissipated. The foundational effort ended ‘mid-air’ around 1940 (Davis and Hersh, 1980, p. 323). As shown by Imre Lakatos (1967), some of the leading contemporary participants in foundational studies, from Russell to John von Neumann, concluded that the foundational program had collapsed. For instance, in 1947 von Neumann concluded that ‘Hilbert’s program is essentially dead’ and that ‘it is hardly possible to believe in the existence of an absolute immutable concept of mathematical rigour dissociated from all human experience’ (von Neumann, 1947). It would of course be false to interpret this as an indication that Wittgenstein’s critique of the foundational program has been accepted, for that is surely not the general situation.112 Other developments have played a larger part. 112 In fact, Wittgenstein’s manuscripts on mathematics were met with sometimes excited opposition, at first at least (key texts from this debate can be found in Shanker, 1986a). The situation is now somewhat changed, though (cf., e.g., Shanker, 1987a; Tait, 2005). Chapter 13: The mythology of computational artifacts 425 Writing in 1936, Friedrich Waismann, Wittgenstein’s ally in the early 1930s, observed that, although it had ‘looked as if Hilbert’s methods of attack would lead to the desired end’, the situation ‘changed essentially’ with the publication of Gödel’s article on ‘undecidable propositions’ (1936, pp. 100 f.). In Waismann’s interpretation, Gödel showed that ‘the consistency of a logico-mathematical system can never be demonstrated by the methods of this system’: ‘We had previously visualized mathematics as a system all of whose propositions are necessary consequences of a few assumptions, and in which every problem could be solved by a finite number of operations. The structure of mathematics is not properly rendered by this picture. Actually mathematics is a collection of innumerably many coexisting systems which are mutually closed by the rules of logic, and each of which contains problems not decidable within the system itself.’ ‘Mathematics is not one system but a multitude of systems; we must, so to speak, always begin to construct anew.’ (Waismann, 1936, pp. 102, 120). This picture by a colleague of Wittgenstein may not quite be official doctrine, but it is close. The notion of a universal formal language that could encompass the motley of mathematics and subject them to one uniform representational form has all but evaporated. At any rate, Waismann’s description seems like an accurate description of the proverbial ‘the facts on the ground’, for, as one observer has put it, ‘Like the hordes and horses of some fabulous khan, today’s mathematicians have ridden off in all directions at once, conquering faster than they can send messages home.’ (Bergamini, 1963, p. 169). The ‘motley of mathematics’ has asserted itself. There is also the possibility that mathematicians have simply learned to live without Russell’s high-strung faith in ‘the realm of absolute necessity’ and have learned to subsist and get on with their business without being fortified by Hilbert’s conviction of the guaranteed solvability of mathematical problems. And in this mathematicians may have, at least indirectly, been influenced by Wittgenstein’s persistent attempts to develop a cure against epistemological hypochondria. Anyway, what first of all killed off the foundational program was probably the mundane fact that, as pointed out by Davis and Hersh in their very insightful and balanced account of The Mathematical Experience, the entire program was ‘not compatible with the mode of thought of working mathematicians’, for ‘From the viewpoint of the producer, the axiomatic presentation is secondary. It is only a refinement that is provided after the primary work, the process of mathematical discovery’ (Davis and Hersh, 1980, p. 343). No surprise then that mathematicians are now rather pragmatic about the epistemological issues that previously bewitched the foundationalists. In the words of the notable mathematician Paul Cohen: ‘The Realist [i.e., Platonist] position is probably the one which most mathematicians would prefer to take. It is not until he becomes aware of some of the difficulties in set theory that he would even begin to question it. If these difficulties particularly upset him, he will rush to the shelter of [Hilbert’s] Formalism while his normal position will be somewhere between the two, trying to enjoy the best of two worlds.’ (P. J. Cohen, 1967, p. 11) 426 Cooperative Work and Coordinative Practices Citing this, Davis and Hersh add — rather irreverently — that ‘Most writers on the subject seem to agree that the typical working mathematician is a Platonist on weekdays and a formalist on Sundays. […] The typical mathematician is both a Platonist and a formalist — a secret Platonist with a formalist mask that he puts on when the occasion calls for it’ (Davis and Hersh, 1980, pp. 321 f.). That is, when engaged in doing mathematics mathematicians conceive of themselves as engaged in discovering existent objects and relations. This is, in Schutzian terms, the ‘natural attitude’ of mathematicians: the necessary assumptions of their daily work, the conceptual horizon that is, and has to be, taken for granted to get on with business. But when pressed to argue the status of these elusive objects and relations prior to their discovery, mathematicians will adopt the official policy of formalism: that all they do is really to manipulate (arcane) symbols. The sound and fury of the ‘foundations crisis’ was not in vain, however; not at all. Sure, as far as regaining the former serene certitude of mathematics is concerned, the enormous effort Russell and Whitehead and many others put into the foundational program came to naught, but unlike Babbage’s misguided attempts at ‘unerringly certain’ calculation ‘by steam’, these efforts, while also misguided on Wittgenstein’s view, were not wasted. In their failed attack on the ‘foundations crisis’ the techniques of mathematical logic were developed immensely compared to what had been achieved by Frege and Peano before them, and in doing so they provided the means for the formalization of algorithmics. The notion of an algorithm had, of course, been known and applied for ages (cf., e.g., Chabert, 1999); but what the foundations program accomplished was to provide algorithmics with requisite techniques and thus make algorithm a rigorous concept. The story is of interest to us because the foundationalists, in the course of trying to overcome the ‘foundations crisis’, developed what — incidentally — became essential parts of the conceptual foundation of computing technology. As Russell’s biographer, Ray Monk, puts it: ‘in the process (and this is perhaps where the lastingly important aspect of the work lies), [Russell and Whitehead] had given an enormous boost to the development of mathematical logic itself, inventing techniques and suggesting lines of thought that would provide the inspiration for subsequent mathematical logicians, such as Alan Turing and John von Neumann, whose work, in providing the theoretical basis for the theory of computing, has changed our lives.’ (Monk, 1996, p. 195). When Monk uses a phrase like ‘the theoretical basis for the theory of computing’, a note of caution is required, however. Computing is not a sharply defined technology but rather a family of technologies; nor are computing technologies developed on the basis of a clearly delimited theoretical basis. Like mathematics, computing is a motley of techniques and concepts that defies overarching formalization. However, the techniques and concepts of mathematical logic have been essential to the development of computing technologies. The very concept of computational artifact originated in these ‘metamathematical’ efforts. The metaphysical excesses of ‘metamathematics’ stick to the philosophy of computing. This also means that Wittgenstein’s critique of the way in which this work was understood Chapter 13: The mythology of computational artifacts 427 by those involved is of the highest relevance for the way in which we in CSCW and related fields conceive of computational artifacts and investigate their use in cooperative work practices. 1.2. Turing’s ambiguous machine Turing’s famous article, ‘On computable numbers…’, from 1936 was designed as an attempt to investigate a central problem in Hilbert’s program that had not been addressed by Gödel, namely the problem of decidability. And like Gödel had done with respect to the question of completeness and consistency, Turing demonstrated that no consistent formal system of arithmetic is decidable. Turing began the article by considering the work of a human computer, ‘a man in the process of computing’, Like de Prony and many others before him, he did so in a language that is deliberately mechanistic: ‘We may compare a man in the process of computing a real number to a machine which is only capable of a finite number of conditions q1, q2, …, qR which will be called “m-configurations”. The machine is supplied with a tape […] running through it, and divided into sections (called “squares”) each capable of bearing a “symbol”. At any given moment there is just one square […] which is “in the machine”. We may call this square the “scanned square”. The symbol on the scanned square may be called the “scanned symbol”. The “scanned symbol” is the only one of which the machine is, so to speak, “directly aware”’ (Turing, 1936, p. 231). ‘We may compare a man […] to a machine’. A this stage, Turing was explicitly engaged in a comparison. The mechanistic language was obviously used with some caution, as the use of citation marks and expressions such as ‘so to speak’ suggest. After a lengthy argument describing the technical operations performed by the ‘man in the process of computing a real number’ in these mechanistic terms, Turing returned to his analysis of the work of the human computer, but now in less cautious mechanistic language: ‘The behaviour of the computer at any moment is determined by the symbols which he is observing, and his “state of mind” at that moment. […] Let us imagine the operations performed by the computer to be split up into “simple operations” which are so elementary that it is not easy to imagine them further divided. Every such operation consists of some change of the physical system consisting of the computer and his tape. We know the state of the system if we know the sequence of symbols on the tape, which of these are observed by the computer […], and the state of mind of the computer. We may suppose that in a simple operation not more than one symbol is altered. Any other changes can be split up into simple changes of this kind. The situation in regard to the squares whose symbols may be altered in this way is the same as in regard to the observed squares. We may, therefore, without loss of generality, assume that the squares whose symbols are changed are always “observed” squares.’ (Turing, 1936, p. 250) Turing then summarized the results of his argument so far: ‘The simple operations must therefore include: (a) Changes of the symbol on one of the observed squares. (b) Changes of one of the squares observed to another square within L squares of one of the previously observed squares. 428 Cooperative Work and Coordinative Practices The operation actually performed is determined […] by the state of mind of the computer and the observed symbols. In particular, they determine the state of mind of the computer after the operation is carried out.’ (Turing, 1936, p. 251). Having talked, so far, of a person involved in some kind of computing, Turing now — and rather inconspicuously — slipped into describing ‘a machine to do the work of this computer’ in exactly the same language that was earlier used to characterize the operations of the human computer: ‘We may now construct a machine to do the work of this computer. To each state of mind of the computer corresponds an “m-configuration” of the machine..’ (Turing, 1936, p. 251). It was no longer simply a comparison. Turing now, at the end of his discussion and without further argument, conceptually equated the human computer’s ‘state of mind’ to an ‘m-configuration’, that is, to what today is called a program. Later he would do that without any of the caution he displayed at the beginning of his famous paper. For example, in the lecture he gave to the London Mathematical Society in 1947, the ‘rules of thumb’ of the human computer and the ‘machine process’ are without further ado referred to as ‘synonymous’: ‘One of my conclusions was that the idea of a “rule of thumb” process and a “machine process” were synonymous’ (Turing, 1947, p. 378). What Turing implied in 1936 but made explicit in 1947 was the ‘mechanist thesis’ that underlies the cognitivist concept of rational behavior: the execution of ‘plans’, ‘schemes’, etc. in human practices and in machines are categorially identical. 1.3. Conceptions of the ‘mechanical’ In obvious reference to Turing’s concept of mechanical calculation, Wittgenstein (in a manuscript written 1942-44) raised the startling question: ‘Does a calculating machine calculate?’, and to answer that he introduced a (striking, if hastily sketched) thought experiment: ‘Does a calculating machine calculate? Imagine that a calculating machine had come into existence by accident; now someone accidentally presses its knobs (or an animal walks over it) and it calculates the product 25 × 20. I want to say: it is essential to mathematics that its signs are also employed in mufti. It is the use outside mathematics, and so the meaning of the signs, that makes the sign-game into mathematics.’ (Wittgenstein, 1937-44, V §2) That is, for a calculating machine to be said to calculate it has to be part of a practice in which techniques of calculation are mastered and routinely performed, that is, ‘employed in mufti’, applied to do ordinary work of that sort. Let me elaborate his scenario a little. Imagine some ecologist doing field work somewhere in the jungles of Central Africa. Let us say that he is collecting data on the local ecological system and that he at some point is doing some statistical calculation using one of the $100 computers designed by the MIT Media Lab. Now, for some reason - perhaps while taking a nap — he leaves abruptly, leaving the computer unattended until its battery runs out of power. Shortly afterwards a Chapter 13: The mythology of computational artifacts 429 group of bonobo chimpanzees comes along and finds it. Out of idle curiosity one of them turns the handle, winding up the device that is then suddenly re-activated and continues the operations it was performing before its battery ran out of power. Is the computer still doing calculations? The computer is still the same thing and it still behaves in the usual highly regular manner, but does it still serve as a computer? What Wittgenstein was suggesting with his scenario is that the concept of calculation is of the normative category; it implies that the whoever does the calculation understands the rules of the calculus in question; that is, that the calculator has the ability to apply the rules and can justify the procedure and the result with reference to the rules. However, it is not as clear-cut as this, for we also have the concept of somebody doing calculations ‘mechanically’. Thus, in the following section, Wittgenstein introduced another, quite different, scenario involving a human computer: ‘But is it not true that someone with no idea of the meaning of Russell’s symbols could work over Russell’s proofs? And so could in an important sense test whether they were right or wrong? A human calculating machine might be trained so that when the rules of inferences were shewn it and perhaps exemplified, it read through the proofs of a mathematical system (say that of Russell), and nodded its head after every correctly drawn conclusion, but shook its head at a mistake and stopped calculating. One could imagine this creature as otherwise perfectly imbecile.’ (Wittgenstein, 1937-44, V §3) The question implicitly suggested here is whether the human computer, an ‘otherwise perfectly imbecile’, can in fact be said to be making logical inferences, even when the ‘human calculating machine’ has no idea of the meaning of Russell’s system? The case of the misplaced calculator and that of the ‘perfectly imbecile’ ‘human calculating machine’ point to a serious muddle in our notions of calculation, computing, and mechanical procedures. Let us try to sort out some of the pertinent distinctions: (i) Ordinary (competent) calculation: A human computer has learned the rules of a certain calculus and masters the concomitant methods and techniques. Having learned to follow the rules, the human computer is able to apply the rules appropriately, explain and justify his or her moves with reference to the rules, etc. (ii) Routine calculation: Again, a human computer has learned the rules of a certain calculus and masters the pertinent methods and techniques. However, since the human computer has performed this kind of calculation often, he or she proceeds confidently, unhesitatingly, and without forethought. For that reason this is often referred to as ‘mechanical calculation’ and can of course be said to be ‘mechanical’ in as much as the human computer proceeds steadily and effortlessly, but, as Wittgenstein points out, it is not performed mindlessly: ‘One follows the rule mechanically. Hence one compares it with a mechanism. “Mechanical”? That means: without thinking. But entirely without thinking? Without reflecting.’ (Wittgenstein, 1937-44, VII §60) 430 Cooperative Work and Coordinative Practices The human computer masters the rules but does not necessarily have them present and may need a minute or two to recall the rules, if asked for an explanation, for example. But still, we would find it very strange indeed if the human computer would not be able to recognize a fault if one was found and pointed out to him. (iii) Distributed routine calculation: As exemplified by the manufactures form of cooperative calculation work devised by de Prony, this is the prototypical case of human computing. For a certain task, an algorithm has been specified, and the constituent partial or ‘atomic’ operations have been ‘divided’ and allocated among multiple ‘human computers’, in a carefully calculated proportion and sequence. Although the cooperative effort as a whole may be organized according to the rules of a sophisticated algorithm, the partial operations of the individual human calculator may be extremely rudimentary and, as in de Prony’s example, reduced to filling in a predesigned spreadsheet by operations of addition and subtraction. The work of the human computers may also in this case be characterized as ‘mechanical’; but again, the human computers do not perform their operations thoughtlessly but rather ‘without reflection’, as a matter of course, with a confidence rooted in routine. What is specific is that they, while being able to explain and justify (by invoking the rules) what they do individually, in their ‘atomic’ operations, nevertheless typically will be unable to account for the rules of the algorithm governing the cooperative effort in its totality. This is then often turned into a claim that multitude ‘meaningless’ operations on an aggregate level may appear as complex calculation. But in making such an inference one forgets the algorithm carefully designed by de Prony and his fellow mathematicians. Anyway, there was nothing ‘meaningless’ about the work of the hairdressers: they were conscientiously following the rules of additions and subtraction as well as the protocol embedded in the prefabricated spreadsheet: ‘it makes no sense to speak of a “meaningless sub-rule”, simply because it is unintelligible to speak of following a rule — no matter how simple or complex — without understanding that rule’ (Shanker, 1987c, p. 89). One can compare with someone who has learned to move chess pieces according to the rules governing their movement (pawns move forward, bishops move diagonally, etc.) but, perversely, has never been taught that it is a game and that it is about winning by trapping the opponent’s king. The imbecile player, or rather, chess piece mover, is able to move the pieces about in accordance with the rules but cannot be said to play chess: ‘the game, I should like to say, does not just have rules; it has a point’ (Wittgenstein, 1937-44, I §20, cf. also III §85). Similarly, the hairdressers cannot be said to be calculating trigonometric tables, but one could certainly say that the entire ensemble, that is, the 80 hairdressers in their specific formation together with de Prony and his assistants, as an ensemble, certainly were calculating trigonometric tables. (iv) The machine-as-symbol: Mathematicians conceive of what they do in terms of the ‘necessity’ and ‘inexorability’ of mathematical propositions once ‘proved’ and this notion of ‘necessary’ and ‘inexorable’ inferences is often expressed in metaphors such as ‘machinery’ and ‘mechanical’. There is a age-long Chapter 13: The mythology of computational artifacts 431 tradition for using such metaphors to express the sense that the result of a calculation is somehow already there even before the calculation started. Wittgenstein’s discussion of this is instructive (Wittgenstein, 1939, pp. 196-199). Speaking ‘against the idea of a “logical machinery”’, he is arguing that ‘there is no such thing’. He gives as an example the use of an imaginary mechanism such as an idealized crank shaft in a geometrical proof: Figure 1. Wittgenstein’s example of a ‘kinematic’ proof (Wittgenstein, 1939, p. 195). In this kind of proof one works out how the piston will move if the crankshaft is moved in a particular way, and in doing so, ‘One always assumes that the parts are perfectly rigid. — Now what is this? You might say, “What a queer assumption, since nothing is perfectly rigid.” What is the criterion for rigidity? What do we assume when we assume the parts are rigid?’, Wittgenstein asks and goes on: ‘In kinematics we talk of a connecting rod—not meaning a rod made of steel or brass or what-not. We use the word “connecting rod” in ordinary life, but in kinematics we use it in quite a different way, although we say roughly the same things about it as we say about the real rod: that it goes forward and back, rotates, etc. But then the real rod contracts and expands, we say. What are we to say of this rod: does it contract and expand?—And so we say it can’t. But the truth is that there is no question of it contracting or expanding. It is a picture of a connecting rod, a symbol used in this symbolism for a connecting rod. And in this symbolism there is nothing which corresponds to a contraction or expansion of the connecting rod.’ The machine is not a machine but a symbolic machine that stands as a symbol for a certain mathematical operation: ‘If we talk of a logical machinery, we are using the idea of a machinery to explain a certain thing happening in time. When we think of a logical machinery explaining logical necessity, then we have a peculiar idea of the parts of the logical machinery—an idea which makes logical necessity much more necessary than other kinds of necessity. If we were comparing the logical machinery with the machinery of a watch, one might say that the logical machinery is made of parts which cannot be bent. They are made of infinitely hard material—and so one gets an infinitely hard necessity.’ (Wittgenstein, 1939, p. 196) When mathematicians use terms such as ‘machine’, ‘mechanism’, and ‘mechanical’ they are talking about a ‘picture’, a ‘symbol’: ‘if I say that there is no such thing as the super-rigidity of logic, the real point is to explain where this idea of super-rigidity comes from—to show that the idea of super-rigidity does not come from the same source which the idea of rigidity comes from’. ‘It seems as if we had got hold of a hardness which we have never experienced’. Like the idea of ‘the inexorability or absolute hardness of logic’ the idea of ‘super-hardness’ or ‘super-rigidity’ of the ‘machine-as-symbol’ is a strangely inverted expression of the normative nature of mathematics: the parts of the machine are super-hard be- 432 Cooperative Work and Coordinative Practices cause they are not part of any material machine but are rules we do not question simply because they are rules and are very carefully integrated and connected to and integrated in the grammar of our number words: ‘“It isn’t a machine which might be explored with unexpected results, a machine which might achieve something that couldn’t be read off from it. That is, the way it works is logical, it’s quite different from the way a machine works. Qua thought, it contains nothing more than was put into it. As a machine functioning causally, it might be believed capable of anything; but in logic we get out of it only what we meant by it.”’ (Wittgenstein, 1931-34, p. 247). In sum, Turing’s machine anno 1936 is not a machine but a symbolic machine, that is, a calculus. ‘If calculating looks to us like the action of a machine, it is the human being doing the calculation that is the machine’ (Wittgenstein, 1937-44, IV §20). We accept the rules as incontestable and let our calculations be guided by the rules: by doing so, we rely on the rules to lead us to correct results, just like we rely on the ruler to guide us to draw a straight line. It is this that in mathematical logic is meant by ‘mechanical procedure’. (v) Machine calculation: Now, here ‘machine’ (and ‘mechanical’, etc.) means something categorially different from the mathematician’s ‘machine-as-symbol’, namely, the use of a causal process (classical mechanics, hydraulics, electromechanical, analog electronics, digital electronics, software machines), that has been carefully devised to operate in an extremely regular manner (behaving uniformly, dependably) and thus to transform specific input into specific output according to some pattern with extreme degrees of probability. Although an electronic computer operates in a highly regular manner, it is unintelligible to say that it follows rules when operating, when, for example, performing calculations; it is a transgression of the logical boundaries demarcating grammatically distinct categories to ascribe rules and rule-following to a machine: in short, it is a transgression of sense. What does make sense is to say that we make the machine behave as if it follows rules and perform calculations, in the same way (only much faster) as we might use a mechanical calculator to do the same. Or better, it is we who follow rules by, in part, using mechanical calculators or digital computers. And it is we who use the machine to produce certain transformations of physical patterns that, within a certain domain of practice, can be taken as calculations, representations, letters, plans, etc. and are routinely taken as such. A computational artifact is a computational artifacts only within a practice of ‘computation’ that, like other forms of rule-following, is a normative activity, whereas the behavior of a computational artifact is the manifestation of causal processes. Now, computation by means of electronic computers is certainly different from computation by means of paper and pen, by means of using slide rulers or abacuses, and so on. Computational artifacts are machines: they operate more or less automatically. That is to say, they undergo highly dependable, causally regulated state changes without human intervention (at least for a period of time). But computational artifacts are still a technique of computation by virtue of being used by practitioners in their normatively defined practices. Chapter 13: The mythology of computational artifacts 433 The concept of ‘mechanical’ performance is obviously not a sharply defined concept. It is a concept developed over centuries by millwrights, blacksmiths, machinists, engineers, and — lately — mathematicians and computer scientists.113 In short, it is a concept with a history, and a rather mixed one at that. So it may be warranted to call it a family-resemblance concept. But if so, it is a familyresemblance concept deeply divided by a categorial split, encompassing concepts in disparate grammatical domains. It is noteworthy, however, that our language has a large number of words that we use to make far more fine-grained distinctions, if not always very sharply or consistently: rule formulation in the form of prose, lists, structured English, blueprints, etc.; rule codification, i.e., rules arranged according to a specific scheme; rule formalization, i.e., rules encoded in a notation (a finite set of symbols); software coding, i.e., the formalized rules expressed in the calculus of a programming language; and implementation, i.e., incorporation of the code on a specific platform, taking into account the specific instruction set, word length, memory constraints, etc. That is, it is not so that the distinctions are not made by practitioners but different practitioners of different ilk make the distinctions differently. This is as it always is, and normally causes no harm, as these different uses typically co-exist without intersecting. The problem arises when practitioners — and philosophizing theorists — make general propositions about ‘mechanical’ with no regard for the conceptual multiplicity. In much of the literature on the theory of computation, even among the most notable scholars, these different uses of the notion of ‘mechanical’ are thrown together in the most primitive manner. Take, for example, Hao Wang’s discussion of Turing’s contribution to our understanding of algorithms and mechanical calculation. He started by defining the concept of algorithm as ‘a finite set of rules which tell us, from moment to moment, precisely what to do with regard to a given class of problems’, and then, to illustrate the concept of algorithm, introduced the ubiquitous schoolboy as an example of a human calculator: ‘a schoolboy can learn the Euclidian algorithm correctly without knowing why it gives the desired result’ (Wang, 1974, p. 90). This argument is a central tenet in the philosophy of computing — but also deeply confused. There is, in Shanker’s words, ‘something profoundly out of focus’ in this last sentence: ‘the thoughts which it contains are pulling in opposite directions, crediting the schoolboy with the very piece of knowledge which is immediately denied him’ (Shanker, 1986c, p. 22). Shanker then elaborates his objections: ‘to learn to apply an algorithm correctly involves more than merely producing the “right” results. In order to say of the schoolboy that he has learnt — grasped — the rule, we will demand more than simply the set of his results to justify such a judgment. The criteria for crediting someone with the mastery of a rule are more complex than this; we place it against the background of his explaining, justifying, correcting, answering certain questions, etc. We would no doubt be very puzzled by the case of a schoolboy who could invariably give the “right” result for an algorithm 113 It does not help that the terms ‘machine’ and ‘mechanical’ have been adopted by sociologists as a metaphor to characterize certain organizational forms such as, for example, the cold indifference and imperviousness of state bureaucracies: the ‘mechanical efficiency’ that ‘reduces every worker to a cog in this machine’ (Weber, 1909, p. 127). 434 Cooperative Work and Coordinative Practices and yet could not provide us with absolutely no information about how or why he had derived that result.’ (Shanker, 1986c, p. 22). The source of the confusion is here that Wang, like so many other philosophers of mathematics and logic, confused the mastery of the steps of the algorithm with mastery of the algorithm. In the words of Shanker, ‘All that we could rightly say in such a situation is that the schoolboy has learned a series of (for him) independent rules; but to learn how to apply each of these sub-rules does not amount to learning the algorithm’ (Shanker, 1986c, p. 22). The confusion is so much more remarkable as Wang, immediately after having first credited and then debited the schoolboy with knowledge of the algorithm, commented that: ‘In practice, when a man calculates, he also designs a small algorithm on the way. But to simplify matters, we may say that the designing activity is no longer a part of his calculating activity.’ (Wang, 1974, p. 90). But in ‘simplifying’ matters this way Wang made himself guilty of setting up a thought experiment in which an algorithm is part of the experimental setup — and then surreptitiously removing the algorithm from his interpretation. This is hardly ‘simplification’ but rather obfuscation. Like Turing (and most mathematical logicians in general) Wang simply took for granted that routine (‘noncreative’) calculation is of the same category as a causal process performed by a machine. He did so quite explicitly by saying that what makes a process ‘mechanical’ is that it is ‘capable of being performed by a machine, or by a man in a mechanical (noncreative) way’ (ibid., p. 91). Wang here made the same transgression of logical boundaries as Turing did when he surreptitiously slipped from his ‘mechanical’ account of the human computer’s work to devising ‘a machine to do the work of this computer’ in exactly the same language: ‘What this means is that whenever the operations of a system or organism can be mapped onto an algorithm, there is no logical obstacle to the transfer of that algorithm to the system or organism itself’ (Shanker, 1987b, p. 39). Turing (and Wang and a host of other philosophers of mathematics, logic, and computing) are flickering between the normative concept of following a rule and the associated notions of correctness and justification etc., and the causal concept of machinery and mechanical. Although it ‘is unintelligible to speak of a machine following a rule’ (Shanker, 1987c, p. 89), a ‘non-normative sense of following rules’ has been introduced (Shanker, 1987b, p. 39). The upshot is a transgression of the boundaries of sense which has created an intellectual muddle with detrimental consequences. It is important to emphasize that their fault is not the notion that software machines can perform sophisticated operations and can do so in flexible ways hitherto unthinkable. Of course not: CSCW is predicated on those notions and the technologies that have been built on them. The issue is neither that software machines can be integrated in (normatively defined) practices so as to automatically regulate aspects of action and interaction. Of course not: this is an observable fact. In fact, strictly speaking, the issue is not even that computer scientists and logicians, in the natural attitude of their own habitat, resort to language that, outside of their domain, can be fatally inaccurate, ambiguous, misleading (‘programming lan- Chapter 13: The mythology of computational artifacts 435 guage’, ‘error’, ‘rule’, ‘calculation’, etc.). It is clear that when computer scientists are engaged in programming, discussing programs, debating the complexity of an algorithm — then such usage is of course completely legitimate. It is simply a kind of shorthand. However, just as the ‘talk of mathematicians becomes absurd when they leave mathematics’ (Wittgenstein, 1932-33, p. 225), the issue arises when computer scientists (and journalists!) bring these professional metaphors into ordinary language. Then a bonfire of metaphysics erupts. The issue is that they, having ventured outside of their respective fields and started philosophizing about ‘a machine following a rule’, degenerate to sheer unintelligible nonsense. This is the real issue. By transgressing the logical boundaries governing the use of concepts of disparate grammatical domains, Turing and his followers have created a conceptual imbroglio in which investigations of computing technologies in and for actual work practices are gravely handicapped. This concept of computational artifact, as received from the philosophy of mathematical logic, is confused and this leads us to ascribe causal characteristics to normative actions while others respond that actions are not causally determined and hence not rule-governed. This leaves no logical space for CSCW. 2. Room for CSCW The point of all this is not to say that CSCW’s research program is centered on the concept of computation, nor even that CSCW is or should be focused on dispelling the metaphysical fog in which the concepts of computing and computational artifact are still shrouded. The point of all this is rather to clear the construction site of the piles of debris and rubble that prevent us from proceeding and from even seeing the challenge and identifying the possibilities and issues. For CSCW to progress and even exist as an intellectually respectable research area requires that fundamental concepts such as ‘computation’, ‘formalization’, ‘computational artifact’, ‘mechanical procedure’, ‘mechanization’, ‘causal process’, ‘regularity’, ‘rule following’, etc. become surveyable, and, a fortiori, that we developed a differentiated conceptualization of the ways in which such techniques and practices function and develop. Without clarity on these issues CSCW will remain a degenerative research program, caught up in endless conceptual confusion and unprincipled design experiments. The issues of the concept of computation and plans and computational artifacts are critically important but are not focal issues for CSCW. They are inevitable issues, not as the subject matter of CSCW, but in the sense that their clarification is the precondition for creating the ‘logical space’ in which CSCW research is intellectually legitimate. CSCW’s research program is centered on the issues of developing computing technologies that enable workers engaged in cooperative work to regulate their interdependent and yet distributed activities effectively, efficiently, dependably, timely, unhurriedly, with dignity, etc. and accordingly, at the same time, enable 436 Cooperative Work and Coordinative Practices them to control, specify, adapt, modify, etc. the behavior of the regulating computational artifacts. The point of clarifying and emphasizing that ‘it is we who are executing the instructions, albeit with the aid of this sophisticated electro-mechanical tool’ (Shanker, 1987c, p. 82), is to let the buzzing fly out of the bottle — to show that there is no conceptual obstructions (notions of rational action as ‘essentially ad hoc’ or of ‘rule following machinery’) prohibiting CSCW from engaging in the development of coordination technologies: the technologies of focal interest to CSCW. With classic machinery, be it a ‘Spinning Jenny’ or a CNC machining station or a digital photo-editing application like Photoshop, the worker interposes, in Marx’s words, a ‘natural process […] as a means between himself and inorganic nature’. By contrast, with coordination technologies we are interposing a ‘natural process’, not as a means between us and the work piece — but as a regulating causal process into the web of our cooperative work relationships. What characterizes coordination technologies is that causal processes are specifically devised and employed as techniques of coordinative practices. That is, with coordination technologies, coordinative practices, the normative constructs of coordinative conduct such as plans, schedules, procedures, schemes, etc., are enacted and imposed (to some extent and in some respects) by means of causally unfolding processes as embodied by computational artifacts. This is not a paradox or a contradiction; it is simply a tension to be explored and mastered and thereby overcome. The fundamental problem of CSCW is defined by this preliminary tension: the design of (causal) mechanisms for (normative) practices. What is new is that we are beginning to construct devices that we use in our coordinative practices — to regulate our interactions. Probably the first example of this kind of technique is the mechanical clock in the town hall towers of medieval Europe that regulated the life of the town. In the words of Lewis Mumford: ‘the regular striking of the bells brought a new regularity into the life of the workman and the merchant. The bells of the clock tower almost defined urban existence’ (Mumford, 1934, p. 14). The clock provided medieval townspeople with a criterion for determining what was the correct time (e.g., ‘morning mass’, ‘bank hour’) under given circumstances, in the form of a commonly accepted and publicly visible metric of time. And the townspeople would typically let their activities be guided by the hands of the clock, or by the sound of its bell, or they might ignore it, if it was of no consequence or if they for some reason had other plans for the day. Now, as mentioned earlier, the development of railroad operations required time-critical coordination on a large scale which in turn motivated the development and deployment of telegraph communication systems and furthermore the establishment of national time measurement conventions and the ‘galvanic’ system of automatic synchronization of time keepers. Like the town-hall clock, the latter system, the system of automatic synchronization, brought ‘a new regularity’ into the life of not just of townspeople within earshot but the lives of millions, in the UK and beyond. The (normative) standard method of time meas- Chapter 13: The mythology of computational artifacts 437 urement (based on Greenwich time) was made effective on such a large scale by the highly regular periodic dissemination of electric pulses over the telegraph network. It was not the causal propagation of the electric pulse that forced railroad workers, travellers, post office clerks, watchmakers, etc. to synchronize their daily conduct but it was these people who followed the rule of standard time measurement by taking note of and adjusting to the time as read from their electromagnetically regularized timekeepers. Now, both the town hall clock and the nationally synchronized clocks of course represented very simple coordination technologies: the clock has two states; it either sounds or it does not, and more sophisticated designs soon had the number of strikes correspond to the time of day. The coordination technologies we now have available are of course integrated into our work practices in significantly more sophisticated ways. A computational calendar system may issue notifications of up-coming tasks and appointments and may be used for calculating and identifying possible time slots for meetings under consideration, etc. Similarly, workflow management systems are used not only for routing documents, case folders, etc, through a system of division of labor, according to case types, formal roles, schedules, sequential dependencies, etc. but may be used also for keeping track of deadlines and for issuing reminders. The point I want to make is it that the problem with computational artifacts is not a principled one. Just as the ruler does not ‘in any strong sense’ make me, i.e., cause me to, draw a straight line but I who exploit its material properties (stiffness, geometry) to draw what is conventionally considered a straight line, it is we who follow rules by using computational artifacts. There is no conceptual problem in this, and the state of angst should be called off. The problems that we do have with using machines in our rule-based practices is a practical one: it is a problem of cost, namely, the cost (effort, time, resources, risk, etc.) of construction and modification of such artifacts. 3. The practical inexorability of CSCW With electronic stored-program computing (equipped with high-level programming languages, code libraries, editors, compilers, interpreters, and what have you), the construction of machinery has become immensely inexpensive. The cost of modifying such machines is negligible compared to previous technologies. Consequently, over the last couple of decades it has become economically feasible to construct vast machine systems, even global ones. Already in the case of the mechanical machine systems of the 19th century, such as printing presses and process industries, workers cooperated through causal processes and the coordination of their local tasks were to a large extent preordained by the structure of the machinery. But their ability to change the protocols embodied in the machine systems were practically non-existent. At best, the only means of coordination available to them during operations were monitoring of 438 Cooperative Work and Coordinative Practices bodily conduct, shouting, conversations (cf. the description of the hot rolling mill in Popitz, et al., 1957). The stored-program architecture that made computing a practically universal control technology has reduced the cost of constructing machinery. And propelled by steady advances in semiconductor technology since the invention of the integrated circuit and especially the microprocessor, the development of computing technology is steadily eroding the cost of constructing and modifying software machinery. As a result, the very concept of ‘automation’ has been transformed. Until quite recently the concept of automation remained identical to the concept of automation developed by Ure and Marx, namely a concept of a technical system that, with only occasional human intervention, provides virtually immutable operational continuity on a large scale, such as power looms, transfer lines, paper mills, nuclear power plants, oil refineries, etc. (cf., e.g. Diebold, 1952; Piel, et al., 1955; Blauner, 1964; Pollock, 1964; Luke, 1972). With software machinery, this notion of ‘wall-to-wall’ automation has been superseded by the interactive-computing concept of fine-grained and highly reactive automatic control mechanisms (e.g., ‘word wrap’) that can interoperate in flexible ways. Computing technologies has provided the technical means for the integration of myriads of automatic processes. Ironically, this means that large-scale machine systems are becoming ubiquitous. Automatic control mechanisms can be deployed in domains and for activities for which classic ‘wall-to-wall’ automation was not feasible at all, simply because production conditions had been too uncertain or variable for it to have been economically viable. With software control mechanisms, making machines interconnect and form large-scale systems is inexpensive and effortless, compared to connecting by rods and gears, as it can be done by transmitting data or code between software machines. Thus, computing technologies, combined with network technologies, has made the construction of the vast machine systems (ERP systems, CAD/CAM systems, financial transaction systems, mobile phone systems) a viable option. Machine systems are or are becoming a dominant technology in all domains of work: in manufacturing and mining; in transportation by air, sea, rail, or road; in construction; in administrative work such as accounting, trading, banking, and insurance; in retail trade (supermarkets); in hospital work; in architectural design, engineering design, and programming; in experimental science; in newspaper production, radio and television program production and broadcasting, and increasingly also in movie production. Consequently, cooperative work is increasingly carried out in and through the artificial causal processes constituted by computational machine systems and the realm of ordinary work correspondingly characterized by increasingly widely ramified cooperative work relationships. On the other hand, and again by virtue of the stored-program architecture, computing technologies also provide the general basis for constructing, at another level of abstraction, coordination technologies that facilitate fine-grained automatic control of routine interdependencies. Chapter 13: The mythology of computational artifacts 439 Not only does computer technology offer the means of control mechanisms that are physically cleanly separated from the mechanisms of transforming, converting, and transmitting energy, and not only does computing technology offers the means of control mechanisms to interconnect machine systems; it also, by the same token, affords the logical and physical separation of mechanisms for the regulation of interdependent cooperative activities from the mechanisms of controlling the various productive processes, be they processes of fabrication, transportation, etc. or processes of accounting, financial transactions, etc. That is, it affords the emergence of coordination technologies as a distinct class of technologies, separate from those of machine systems, as it is becoming technically and economically feasible for ordinary workers to construct and modify the control systems that mediate and regulate their cooperative activities. The concept of the distinct control mechanism is again crucial. In coordination technologies such as workflow systems, scheduling systems, etc., the computational protocols regulating the coordination of interdependent activities are — can be, or rather: should be — technically separate from the procedures of the ‘first order’ machine systems (calculating aircraft vectors or passenger lists, controlling production machinery, etc.). It is therefore technically feasible to modify the computational coordinative protocols at very low costs. In other words, interactive computing technologies make it technically and economically feasible even for ordinary workers to devise, adapt, modify the computational protocols of their coordinative practices. Like the town hall clock, the computational artifact does not cause practitioners to behave in a certain way. Just like the medieval townspeople might let their activities guide by the hands of the clock, or by the sound of its bell, workers today may let computational artifacts guide their interactions. Or, like the medieval townspeople, workers may disregard the presumptive mechanically issued command and decide that they have reasons — valid reasons — not to act on the signals produced by the artifact. Again, the problem is not conceptual; it is not principled; it is practical. The challenge is to devise technologies that allow ordinary workers to construct and modify computational artifacts for regulating their coordinative practices and to do so with minimal effort. To do this, however, requires that we understand how ordinary coordinative artifacts are being constructed, negotiated, adopted, amended, combined and recombined. Is there a pattern, a logic even, to the way they are constructed, put together, combined and recombined, and so on? To address those questions systematically requires that we overcome our intellectual paralysis. One may think of CSCW as a fad; it certainly has had its share of folly. And one may find that it has had it time. But the research problem that prompted the formation of CSCW 25 years ago, was not an arbitrary construction, incidentally devised; it is a problem that emerges again and again in the practical development and application of software machinery in cooperative work settings, in the design of workflow systems and ‘business-process execution languages’, in building computational taxonomies or ‘ontologies’ for scientific work, in building models 440 Cooperative Work and Coordinative Practices of product families, etc. It arises from the issues involved in interposing causal processes as regulatory mechanisms in and for cooperative work relations. That is, the CSCW research program arises from the organization of modern work, from the incessant attempts to make the coordination of interdependent activities more efficient, reliable, flexible, etc. by means of computational artifacts. 441 References Abbate, Janet Ellen: Inventing the Internet, The MIT Press, Cambridge, Mass., and London, 1999. (Paperback ed., 2000). Abbott, Kenneth R.; and Sunil K. Sarin: ‘Experiences with workflow management: Issues for the next generation’, in J. B. Smith; F. D. Smith; and T. W. Malone (eds.): CSCW’94: Proceedings of the Conference on Computer-Supported Cooperative Work, 24-26 October 1994, Chapel Hill, North Carolina, ACM Press, New York, 1994, pp. 113-120. Académie Royale des Sciences: Histoire de l'Académie royale des sciences, M.DC.XCIX [1699], avec les mémoires de mathématique & de physique, tirez des registres de cette Académie, vol. 1, Boudot, Paris, 1702. – Gabriel Martin etc., Paris, 1732. Académie Royale des Sciences: Histoire de l'Académie royale des sciences depuis son etablissement en 1666 jusqu’en 1699, vol. 1-13, Compagnie des libraires, Paris, 1729-34. – Gabriel Martin, etc., Paris, 1732. Académie Royale des Sciences: L'art du meunier, du boulanger, du vermicellier. Text ed. by J.-E. Bertrand. Société typographique. Neuchatel, 1761. (Descriptions des arts et métiers, faites ou approuvées par Messieurs de l'Académie royale des sciences de Paris, vol. 1). (2nd ed., 1771). Académie Royale des Sciences: Descriptions des arts et métiers, faites ou approuvées par Messieurs de l'Académie royale des sciences de Paris, vol. 1-20, Société typographique, Neuchatel, 1771-83. Text ed. by J.-E. Bertrand. (Nouvelle édition publiée avec des observations, & augmentée de tout ce qui a été écrit de mieux sur ces matières, an Allemagne, en Angleterre, en Suisse, en Italie). Ackerman, Mark S.; and Thomas W. Malone: ‘Answer Garden: A tool for growing organizational memory’, in: Proceedings of the Conference on Office Information Systems, April 1990, Cambridge, Mass., ACM, New York, 1990, pp. 31-39. Ackerman, Mark S.: ‘Augmenting the organizational memory: A field study of Answer Garden’, in J. B. Smith; F. D. Smith; and T. W. Malone (eds.): CSCW’94: Proceedings of the Conference on Computer-Supported Cooperative Work, 24-26 October 1994, Chapel Hill, North Carolina, ACM Press, New York, 1994, pp. 243-252. Agre, Philip E.: [Review of] L. A. Suchman: Plans and Situated Actions (Cambridge University Press, Cambridge, 1987), Artificial Intelligence, vol. 43, 1990, pp. 369-384. Agricola, Georgius: De Re Metallica, 1556. Transl. from De Re Metallica, Froben 1556. Transl. by H. C. Hoover and L. H. Hoover. – Dover Publications, New York, 1950. Allport, Alan: ‘Attention and control: Have we been asking the wrong questions? A critical review of twenty-five years’, in D. E. Meyer and S. Kornblum (eds.): Attention and Performance XIV: Synergies in Experimental Psychology, Artificial Intelligence, and Cognitive Neuroscience, MIT Press, Cambridge, Mass., 1993, pp. 183-218. Amber, George H.; and Paul S. Amber: Anatomy of Automation, Prentice-Hall, Englewood-Cliffs, New Jersey, 1962. Andersen, Hans H. K.: ‘Classification schemes: Supporting articulation work in technical documentation’, in H. Albrechtsen (ed.): ISKO’94: Knowledge Organisation and Quality Management, 21-24 June 1994, Copenhagen, Denmark, 1994a. Andersen, Hans H. K.: ‘The construction note’, in K. Schmidt (ed.): Social Mechanisms of Interaction, Computing Department, Lancaster University, Lancaster, UK, September 1994b, pp. 161-186. COMIC Deliverable 3.2. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Andersen, Hans H. K.: Cooperative Documentation Production in Engineering Design: The ‘Mechanism of Interaction’ Perspective, PhD diss., Risø National Laboratory, Roskilde, Denmark, June 1997. 442 Cooperative Work and Coordinative Practices Andersen, N. E., et al.: Professional Systems Development: Experience, Ideas, and Action, Prentice-Hall, Englewood-Cliffs, New Jersey, 1990. Anderson, Robert J.; Wes W. Sharrock; and John A. Hughes: ‘The division of labour’, in: Action Analysis and Conversation Analysis, Maison des Sciences de l’Homme, September 1987, Paris, 1987. Anderson, Robert J.: ‘Representations and requirements: The value of ethnography in system design’, Human-Computer Interaction, vol. 9. Hillsdale, New York, 1994, pp. 151-182. Anderson, Robert J.: ‘Work, ethnography, and system design’, in A. Kent and J. G. Williams (eds.): The Encyclopedia of Microcomputing, vol. 20, Marcel Dekker, New York, 1997, pp. 159-183. Aoki, Masahiko: A New Paradigm of Work Organization: The Japanese Experience, vol. 36, World Institute for Development Economics Research, Helsinki, Finland, 1988. Appelt, Wolfgang: ‘WWW based collaboration with the BSCW system’, Lecture Notes in Computer Science, vol. 1725, 1999, pp. 66-78. Appelt, Wolfgang: ‘What groupware functionality do users really use? Analysis of the usage of the BSCW system’, in K. Klöckner (ed.): PDP 2001: Proceedings of the Ninth Euromicro Workshop on Parallel and Distributed Processing, 7-9 February 2001, Mantova, Italy, IEEE Computer Society Press, 2001, pp. 337-341. Arden, Bruce W. (ed.): What Can be Automated? The Computer Science and Engineering Research Study (COSERS), The MIT Press, Cambridge, Mass., and London, 1980. Aristotle: Nicomachean Ethics (c. 334-322 BCE-a). Transl. by W. D. Ross and J. O. Urmson. In J. Barnes (ed.): The Complete Works of Aristotle. Princeton University Press, Princeton, New Jersey, 1984, vol. 2, pp. 1729-1867. Aristotle: Metaphysics (c. 334-322 BCE-b). Transl. by W. D. Ross. In J. Barnes (ed.): The Complete Works of Aristotle. Princeton University Press, Princeton, New Jersey, 1984, vol. 2, pp. 1552-1728. Aristotle: Politics (c. 334-322 BCE-c). Transl. by B. Jowett. In J. Barnes (ed.): The Complete Works of Aristotle. Princeton University Press, Princeton, New Jersey, 1984, vol. 2, pp. 19862129. Aspray, William: ‘The Institute for Advanced Study computer: A case study in the application of concepts from the history of technology’, in R. Rojas and U. Hashagen (eds.): The First Computers: History and Architectures, The MIT Press, Cambridge, Mass., and London, 2000, pp. 179-194. Atkins, Daniel E., et al.: Revolutionizing Science and Engineering Through Cyberinfrastructure: Report of the National Science Foundation Blue-Ribbon Advisory Panel on Cyberinfrastructure, National Science Foundation, Washington, D.C., January 2003. <http://www.nsf.gov/cise/sci/reports/atkins.pdf> Augarten, Stan: Bit by Bit: An Illustrated History of Computers, George Allen & Unwin, London, 1984. Austin, John Longshaw: Philosophical Papers, Oxford University Press, Oxford, 1961. Text ed. by J. O. Urmson and G. J. Warnock. (3rd ed., 1979). Babbage, Charles: A Letter to Sir Humphry Davy, Bart., President of The Royal Society, etc. etc. on the Application of Machinery to the Purpose of Calculating and Printing Mathematical Tables (London, 3 July 1822). Text ed. by P. Morrison and E. Morrison. In C. Babbage: On The Principles and Development of the Calculator, and Other Seminal Writings by Charles Babbage and Others. Dover Publications, New York, 1961, pp. 298-305. Babbage, Charles: On the Economy of Machinery and Manufactures, Charles Knight, London, 1832. (4th ed., 1835). – Augustus M. Kelley, Publishers, Fairfield, New Jersey, 1986. Babbage, Charles: Passages from the Life of a Philosopher, Longman, etc., London, 1864. Text ed. by M. Campbell-Kelly. – William Pickering, London, 1994. Bacon, Frances: The New Organon: Or True Directions Concerning the Interpretation of Nature (1620). Transl. from Novum Organum, Sive Vera de Interpretatione Naturae. Transl. by J. Spedding; R. L. Ellis; and D. D. Heath. In F. Bacon: The Works. Taggard and Thompson, Boston, 1863. Bahrdt, Hans Paul: Industriebürokratie: Versuch einer Soziologie des Industrialisierten Bürobetriebes und seiner Angestellten, Ferdinand Enke Verlag, Stuttgart, 1958. References 443 Bahrdt, Hans Paul: ‘Die Krise der Hierarchie im Wandel der Kooperationsformen’, in: Verhandlungen des 14. deutschen Soziologentages, Stuttgart 1959, Deutsche Gesellschaft für Soziologie, 1959, pp. 113-121. – [Reprinted in Renate Mayntz (ed.): Bürokratische Organisation, Kiepenheuer & Witsch, Köln-Berlin, 1971, pp. 127-134.]. Baker, Gordon P.; and Peter M. S. Hacker: Scepticism, Rules and Language, Basil Blackwell Publisher, Oxford, 1984. Baker, Gordon P.; and Peter M. S. Hacker: Wittgenstein: Rules, Grammar, and Necessity: Essays and Exegesis of §§ 185-242. Basil Blackwell. Oxford, 2009. (G. P. Baker and P. M. S. Hacker: An Analytical Commentary on the Philosophical Investigations, vol. 2). (2nd rev. ed.; 1st ed.1985). Bannon, Liam J.: ‘Computer-Mediated Communication’, in D. A. Norman and S. W. Draper (eds.): User Centered System Design, Lawrence Erlbaum, New Jersey, 1986, pp. 433-452. Bannon, Liam J.; Niels Bjørn-Andersen; and Benedicte Due-Thomsen: ‘Computer Support for Cooperative Work: An appraisal and critique’, in: Eurinfo’88: 1st European Conference on Information Technology for Organisational Systems, 16-20 May 1988, Athens, Greece, 1988. Bannon, Liam J.; and Kjeld Schmidt: ‘CSCW: Four characters in search of a context’, in P. Wilson; J. M. Bowers; and S. D. Benford (eds.): ECSCW’89: Proceedings of the 1st European Conference on Computer Supported Cooperative Work, 13-15 September 1989, Gatwick, London, London, 1989, pp. 358-372. Bannon, Liam J.; and Susanne Bødker: ‘Beyond the interface: Encountering artifacs in use’, in J. M. Carroll (ed.): Designing Interaction. Psychology at the Human-Computer Interface, Cambridge University Press, Cambridge, 1991, pp. 227-253. Bannon, Liam J.; and Susanne Bødker: ‘Constructing common information spaces’, in J. A. Hughes, et al. (eds.): ECSCW’97: Proceedings of the Fifth European Conference on Computer-Supported Cooperative Work, 7–11 September 1997, Lancaster, U.K., Kluwer Academic Publishers, Dordrecht, 1997, pp. 81-96. Barber, Gerald R.; and Carl Hewitt: ‘Foundations for office semantics’, in N. Naffah (ed.): Office Information Systems, INRIA/North-Holland, Amsterdam, 1982, pp. 363-382. Barber, Gerald R.: ‘Supporting organizational problem solving with a work station’, ACM Transactions on Office Information Systems, vol. 1, no. 1, January 1983, pp. 45-67. Barber, Gerald R.; Peter de Jong; and Carl Hewitt: ‘Semantic support for work in organizations’, in R. E. A. Mason (ed.): Information Processing ’83: Proceedings of the IFIP 9th World Computer Congress, 19-23 September 1983, Paris, France, North-Holland, Amsterdam, 1983, pp. 561-566. Bardini, Thierry: Bootstrapping: Douglas Engelbart, Coevolution, and the Origins of Personal Computing, Stanford Research Institute, Stanford, Calif., 2000. Bardram, Jakob E.: ‘Plans as situated action: An activity theory approach to workflow systems’, in J. A. Hughes, et al. (eds.): ECSCW’97: Proceedings of the Fifth European Conference on Computer-Supported Cooperative Work, 7–11 September 1997, Lancaster, U.K., Kluwer Academic Publishers, Dordrecht, 1997, pp. 17-32. Barnard, Chester I.: The Functions of the Executive, Harvard University Press, Cambridge, Mass., 1938. Barnes, Susan B.: Computer-Mediated Communication: Human-to-Human Communication Across the Internet, Allyn and Bacon, Boston, etc., 2003. Basalla, George: The Evolution of Technology, Cambridge University Press, Cambridge, 1988. Baum, Claude: The System Builders: The Story of SDC, System Development Corporation, Santa Monica, Calif., 1981. Beckert, Manfred: Johann Beckmann, B. G. Teubner, Leipzig, 1983. Beckmann, Johann: Politisch-Oekonomische Bibliothek worinn von den neusten Büchern, welche die Naturgeschichte, Naturlehre und die Land- und Stadtwithschaft betreffen, zuverlässige und vollständige Nachrichten ertheilt werden, Vandenhoeck und Ruprecht, Göttingen, 1772. Beckmann, Johann: Anleitung zur Technologie, oder zur Kenntniß der Handwerke, Fabriken und Manufacturen, vornehmlich derer, welche mit der Landwirthschaft, Polizey und Cameralwissenschaft in nächster Verbindung stehn; nebst Beyträgen zur Kunstgeschichte, Vandenhoeck und Ruprecht, Göttingen, 1777. (5th ed., 1802). 444 Cooperative Work and Coordinative Practices Bell, C. Gordon: ‘Toward a history of (personal) workstations’, in A. Goldberg (ed.): A History of Personal Workstations, Addison-Wesley, Reading, Mass., 1988, pp. 4-47. Benford, Steven D.: ‘A data model for the AMIGO communication environment’, in R. Speth (ed.): EUTECO ’88: European Teleinformatics Conference on Research into Networks and Distributed Applications, 20-22 April 1988, Vienna, Austria. Organized by the European Action in Teleinformatics COST 11ter, North-Holland, Brussels and Luxemburg, 1988, pp. 97109. Benford, Steven D.; and Lennart Fahlén: ‘A spatial model of interaction large virtual environments’, in G. De Michelis; C. Simone; and K. Schmidt (eds.): ECSCW’93: Proceedings of the Third European Conference on Computer-Supported Cooperative Work, 13-17 September 1993, Milano, Italy, Kluwer Academic Publishers, Dordrecht, 1993, pp. 109-124. Benford, Steven D., et al.: ‘Managing mutual awareness in collaborative virtual environments.’, in G. Singh; S. K. Feiner; and D. Thalmann (eds.): VRST’94: Proceedings of the ACM SIGCHI Conference on Virtual Reality and Technology, Singapore, 23-26 August 1994, ACM Press, New York, 1994a, pp. 223-236. Benford, Steven D.; Lennart E. Fahlén; and John M. Bowers: ‘Supporting social communication skills in multi-actor artificial realities’, in: ICAT’94: Proceedings of Fourth International Conference on Artificial Reality and Tele-Existence, 14-15 July 1994, Nikkei Hall, Otemachi, Tokyo, Japan, 1994b, pp. 205-228. <ic-at/papers/list.html> Benford, Steven D.; and Chris Greenhalgh: ‘Introducing third party objects into the spatial model of interaction’, in J. A. Hughes, et al. (eds.): ECSCW’97: Proceedings of the Fifth European Conference on Computer-Supported Cooperative Work, 7–11 September 1997, Lancaster, U.K., Kluwer Academic Publishers, Dordrecht, 1997, pp. 189-204. Bennett, Stuart: A History of Control Engineering, 1800-1930, Peter Peregrinus, London, 1979. (Paperback ed., 1986). Bennett, Stuart: A History of Control Engineering, 1930-1955, Peter Peregrinus, London, 1993. Bentley, Richard; and Paul Dourish: ‘Medium versus mechanism: Supporting collaboration through customization’, in H. Marmolin; Y. Sundblad; and K. Schmidt (eds.): ECSCW’95: Proceedings of the Fourth European Conference on Computer-Supported Cooperative Work, 10–14 September 1995, Stockholm, Sweden, Kluwer Academic Publishers, Dordrecht, 1995, pp. 133-148. Bentley, Richard, et al.: ‘Basic support for cooperative work on the World Wide Web’, International Journal of Human-Computer Studies, vol. 46, 1997a, pp. 827-846. Bentley, Richard; Thilo Horstmann; and Jonathan Trevor: ‘The World Wide Web as enabling technology for CSCW: The case of BSCW’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 6, no. 2-3, 1997b, pp. 111-134. Bergamini, David: Mathematics, Time Inc., New York, 1963. Berners-Lee, Tim: Information management: A proposal, Technical report, CERN, Geneva, May 1990. (2nd ed.; 1st ed. March 1989). – Reprinted in Tim Berners-Lee: Weaving the Web: The Past, Present and Future of the World Wide Web by its Inventor, London and New York, 1999. (Paperback ed., 2000), pp. 229-251. Bertalanffy, Ludwig von: General System Theory: Foundations, Development, Applications, Braziller, New York, 1968. – Penguin Books, Harmondsworth, 1973. Best, Michael H.: The New Competition: Institutions of Industrial Restructuring, Polity Press, Cambridge, 1990. Bewley, William L., et al.: ‘Human factors testing in the design of Xerox's 8010 “Star” office workstation’, in R. N. Smith; R. W. Pew; and A. Janda (eds.): CHI’83: Proceedings of the SIGCHI conference on Human Factors in Computing Systems, 12-15 December 1983, Boston, Mass., ACM Press, New York, 1983, pp. 72-77. Bien, Günther: ‘Das Theorie-Praxis-Problem und die politische Philosophie bei Platon und Aristoteles’, Philosophisches Jahrbuch, vol. 76, 1968-69, pp. 264-314. Bien, Günther: ‘Praxis (Antike)’, in J. Ritter and K. Gründer (eds.): Historisches Wörterbuch der Philosophie, vol. 7, 1989, pp. 1277–1287. References 445 Bignoli, Celsina; and Carla Simone: ‘AI Techniques for supporting human to human communication in CHAOS’, in P. Wilson; J. M. Bowers; and S. D. Benford (eds.): ECSCW’89: Proceedings of the 1st European Conference on Computer Supported Cooperative Work, 13-15 September 1989, Gatwick, London, London, 1989, pp. 133-147. Bittner, Egon: ‘The concept of organization’, Social Research, vol. 32, no. 3, Autumn 1965, pp. 239-255. Bittner, Egon: ‘Objectivity and realism in sociology’, in G. Psathas (ed.): Phenomenological Sociology: Issues and Applications, John Wiley & Sons, New York, etc., 1973, pp. 109-125. Bjerknes, Gro; Pelle Ehn; and Morten Kyng (eds.): Computers and Democracy : A Scandinavian Challenge, Avebury, Aldershot, 1987. Blauner, Robert: Alienation and Freedom: The Factory Worker and His Industry, University of Chicago Press, Chicago, 1964. Bly, Sara; Steve R. Harrison; and Susan Irwin: ‘Media spaces: Bringing people together in a video, audio, and computing environment’, Communications of the ACM, vol. 36, no. 1, January 1993, pp. 28-47. Bogia, Douglas P., et al.: ‘Supporting dynamic interdependencies among collaborative activities’, in S. M. Kaplan (ed.): COOCS’93: Conference on Organizational Computing Systems, 1-4 November 1993, Milpitas, California, ACM Press, New York, 1993, pp. 108-118. Bogia, Douglas P., et al.: ‘Issues in the design of collaborative systems: Lessons from ConversationBuilder’, in D. Z. Shapiro; M. Tauber; and R. Traünmuller (eds.): The Design of Computer Supported Cooperative Work and Groupware Systems, North-Holland Elsevier, Amsterdam, 1996, pp. 401-422. – Originally published at the Workshop on CSCW Design, Schärding, Austria, 1-3 June 1993. Borges, Jorge Luis: ‘On exactitude in science’ (Los Anales de Buenos Aires, 1946). Transl. from ‘Del rigor en la ciencia’. Transl. by N. T. di Giovanni. In J. L. Borges: A Universal History of Infamy. Allen Lane, London, 1973, p. 149. Bourdieu, Pierre: Le sens pratique, Les Éditions de Minuit, Paris, 1980. – English translation: The Logic of Practice, translated by Richard Nice, Polity Press, Cambridge, 1990. Bourdieu, Pierre: The Logic of Practice, Polity Press, Cambridge, 1990. Transl. by R. Nice. (French original, Le sens pratique, 1980). Bowers, John M.; John Churcher; and Tim Roberts: ‘Structuring computer-mediated communication in COSMOS’, in R. Speth (ed.): EUTECO ’88: European Teleinformatics Conference on Research into Networks and Distributed Applications, 20-22 April 1988, Vienna, Austria. Organized by the European Action in Teleinformatics COST 11ter, NorthHolland, Brussels and Luxemburg, 1988, pp. 195-209. Bowers, John M.; and John Churcher: ‘Local and global structuring of computer-mediated communication’, in I. Greif and L. A. Suchman (eds.): CSCW’88: Proceedings of the Conference on Computer-Supported Cooperative Work, 26-28 September 1988, Portland, Oregon, ACM Press, New York, 1988, pp. 125-139. Bowers, John M.: ‘The Janus faces of design: Some critical questions for CSCW’, in J. M. Bowers and S. D. Benford (eds.): Studies in Computer Supported Cooperative Work: Theory, Practice and Design, North-Holland, Amsterdam, 1991, pp. 333-350. Bowers, John M.: ‘The politics of formalism’, in M. Lea (ed.): Contexts of Computer-Mediated Communication, Harvester, New York, 1992, pp. 232-261. Bowers, John M.; Graham Button; and Wes W. Sharrock: ‘Workflow from within and without: Technology and cooperative work on the print industry shopfloor’, in H. Marmolin; Y. Sundblad; and K. Schmidt (eds.): ECSCW’95: Proceedings of the Fourth European Conference on Computer-Supported Cooperative Work, 10–14 September 1995, Stockholm, Sweden, Kluwer Academic Publishers, Dordrecht, 1995, pp. 51-66. Bowker, Geoffrey C.; and Susan Leigh Star: ‘Situations vs. standards in long-term, wide-scale decision-making: The case of the International Classification of Diseases’, in J. F. Nunamaker, Jr. and R. H. Sprague, Jr. (eds.): Proceedings of the Twenty-Fourth Annual Hawaii International Conference on System Sciences, 7-11 January 1991 Kauai, Hawaii, IEEE Computer Society Press, vol. 4, 1991, pp. 73-81. Bowker, Geoffrey C.; and Susan Leigh Star: Sorting Things Out: Classification and Its Consequences, MIT Press, Cambridge, Mass., 1999. 446 Cooperative Work and Coordinative Practices Bradley, Margaret: A Career Biography of Gaspard Clair François Marie Riche de Prony, Bridge Builder, Educator and Scientist, The Edwin Mellen Press, Lewiston, New York, etc., 1998. Brannigan, Augustine: The Rise and Fall of Social Psychology: The Use and Misuse of the Experimental Method, Aldine De Gruyter, New York, 2004. Braudel, Fernand: Civilization and Capitalism, 15th—18th Century. Volume II. The Wheels of Commerce, Harper & Row, New York, 1982. Transl. from Les jeux de l’echange, Librairie Armand Collin, Paris, 1979. Transl. by S. Reynolds. Braun, Ernest; and Stuart Macdonald: Revolution in Miniature: The History and Impact of Semiconductor Electronics, Cambridge University Press, Cambridge, 1978. (Paperback ed., 1980; 2nd ed., 1982). Bromley, Allan G.: ‘Difference and analytical engines’, in W. Aspray (ed.): Computing Before Computers, Iowa State University Press, Ames, Iowa, 1990, pp. 59-98. Brooks, Frederick P., Jr.: ‘No silver bullet: Essence and accidents of software engineering’ (Computer, 1987). In F. P. Brooks, Jr.: The Mythical Man-Month: Essays on Software Engineering. Addison-Wesley, Reading, Mass., 1995, pp. 177-203. Bucciarelli, Louis L.: ‘Reflective practice in engineering design’, Design Studies, vol. 5, no. 3, July 1984, pp. 185-190. Bucciarelli, Louis L.: ‘An ethnographic perspective on engineering design’, Design Studies, vol. 9, no. 3, July 1988a, pp. 159-168. Bucciarelli, Louis L.: ‘Engineering design process’, in F. A. Dubinskas (ed.): Making Time: Ethnographies of High-Technology Organizations, Temple University Press, Philadelphia, 1988b, pp. 92-122. Button, Graham; and Richard H. R. Harper: ‘The relevance of “work practice” for design’, Computer Supported Cooperative Work (CSCW): An International Journal, vol. 4, no. 4, 1995, pp. 263-280. Cabitza, Federico; Marcello Sarini; and Carla Simone: ‘Providing awareness through situated process maps: the hospital care case’, in T. Gross, et al. (eds.): GROUP 2007: International Conference on Supporting Group Work, 4-7 November 2007, Sanibel Island, Florida, USA, ACM Press, New York, 2007, pp. 41-50. Cabitza, Federico; and Carla Simone: ‘“…and do it the usual way”: fostering awareness of work conventions in document-mediated collaboration’, in L. J. Bannon, et al. (eds.): ECSCW 2007: Proceedings of the Tenth European Conference on Computer Supported Cooperative Work, 24-28 September 2005, Limerick, Ireland, Springer, London, 2007, pp. 119-138. Cabitza, Federico; Carla Simone; and Marcello Sarini: ‘Leveraging coordinative conventions to promote collaboration awareness: The case of clinical records’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, August 2009, pp. 301330. Campbell-Kelly, Martin: ‘Punched-card machinery’, in W. Aspray (ed.): Computing Before Computers, Iowa State University Press, Ames, Iowa, 1990, pp. 122-155. Campbell-Kelly, Martin: ‘The Railway Clearing House and Victorian data processing’, in L. BudFrierman (ed.): Information Acumen: The Understanding and Use of Knowledge in Modern Business, Routledge, London and New York, 1994a, pp. 51-74. Campbell-Kelly, Martin: ‘Introduction’. In C. Babbage: Passages from the Life of a Philosopher, William Pickering, London, 1994b, pp. 7-36. Campbell-Kelly, Martin; and William Aspray: Computer: A History of the Information Machine, Basic Books, New York, 1996. Campbell-Kelly, Martin: From Airline Reservations to Sonic the Hedgehog: A History of Software Industry, The MIT Press, Cambridge, Mass., and London, 2003. Card, Stuart K.: ‘The human, the computer, the task, and their interaction: Analytic models and use-centered design’, in D. Steier and T. M. Mitchell (eds.): Mind Matters: A Tribute to Allen Newell, Lawrence Erlbaum, Mahwah, New Jersey, 1996, pp. 259-312. Cardwell, Donald: The Fontana History of Technology, FontanaPress, London, 1994. Carey, Alex: ‘The Hawthorne studies: A radical criticism’, American Sociological Review, vol. 32, no. 3, June 1967, pp. 403-416. References 447 Carroll, John M.: ‘Introduction: The Kittle House Manifesto’, in J. M. Carroll (ed.): Designing Interaction: Psychology at the Human-Computer Interface, Cambridge University Press, Cambridge, 1991, pp. 1-16. Carroll, John M.; Wendy A. Kellogg; and Mary Beth Rosson: ‘The task-artifact cycle’, in J. M. Carroll (ed.): Designing Interaction: Psychology at the Human-Computer Interface, Cambridge University Press, Cambridge, 1991, pp. 74-102. Carroll, John M.: ‘Introduction: Human-computer interaction, the past and the present’, in J. M. Carroll (ed.): Human-Computer Interaction in the New Millennium, ACM Press, New York, 2002, pp. xxvii-xxxvii. Carstensen, Peter H.: ‘The bug report form’, in K. Schmidt (ed.): Social Mechanisms of Interaction, Computing Department, Lancaster University, Lancaster, UK, September 1994, pp. 187-219. COMIC Deliverable D3.2. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Carstensen, Peter H.; Carsten Sørensen; and Henrik Borstrøm: ‘Two is fine, four is a mess: Reducing complexity of articulation work in manufacturing’, in: COOP’95: International Workshop on the Design of Cooperative Systems, 25–27 January 1995, Antibes-Juan-les-Pins, France, INRIA Sophia Antipolis, France, 1995a, pp. 314-333. Carstensen, Peter H.; Carsten Sørensen; and Tuomo Tuikka: ‘Let’s talk about bugs!’, Scandinavian Journal of Information Systems, vol. 7, no. 1, 1995b, pp. 33-54. Carstensen, Peter H.: Computer Supported Coordination, Risø National Laboratory, Roskilde, Denmark, 1996. – Risø-R-890(EN). Carstensen, Peter H.; and Carsten Sørensen: ‘From the social to the systematic: Mechanisms supporting coordination in design’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 5, no. 4, 1996, pp. 387-413. Carstensen, Peter H.; Kjeld Schmidt; and Uffe Kock Wiil: ‘Supporting shop floor intelligence: A CSCW approach to production planning and control in flexible manufacturing’, in S. C. Hayne (ed.): GROUP’99: International Conference on Supporting Group Work, 14-17 November 1999, Phoenix, Arizona, ACM Press, New York, 1999, pp. 111-120. Carstensen, Peter H., et al.: IT-støtte til produktionsgrupper: Debatoplæg fra IDAK-projektet, COIndustri etc., København, June 2001. <http://cscw.dk/schmidt/papers/idak_prod.gr.pdf> Carstensen, Peter H.; and Kjeld Schmidt: ‘Self-governing production groups: Towards requirements for IT support’, in V. Marík; L. M. Camarinha-Matos; and H. Afsarmanesh (eds.): Knowledge and Technology Integration in Production and Services: Balancing Knowledge and Technology in Product and Service Life Cycle. IFIP TC5/WG5.3. Fifth IFIP International Conference on Information Technology in Manufacturing and Services (BASYS’02), 25-27 September 2002, Cancun, Mexico, Kluwer Academic Publishers, Boston, Mass., 2002, pp. 49-60. Ceruzzi, Paul E.: ‘Electronic calculators’, in W. Aspray (ed.): Computing Before Computers, Iowa State University Press, Ames, Iowa, 1990, pp. 223-250. Chabert, Jean-Luc (ed.): A History of Algorithms: From the Pebble to the Microchip, SpringerVerlag, Berlin, Heidelberg, New York, 1999. Transl. from Histoire d’algorithmes. Du caillou á la puce, Paris, 1994. Transl. by C. Weeks. Chalmers, Matthew: ‘Awareness, representation and interpretation’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 11, no. 3-4, 2002. Chambers, Ephraim: Cyclopædia, or, an Universal Dictionary of Arts and Sciences [etc.], J. and J. Knapton, London, 1728. <http://digital.library.wisc.edu/1711.dl/HistSciTech.Cyclopaedia02> Chapman, Robert L., et al.: ‘The Systems Research Laboratory's air defense experiments’, Management Science, vol. 5, no. 3, April 1959, pp. 250-269. Child, John: ‘Organizational design for advanced manufacturing technology’, in T. D. Wall; C. W. Clegg; and N. J. Kemp (eds.): The Human Side of Advanced Manufacturing Technology, John Wiley, Chichester, 1987, pp. 101-133. Ciborra, Claudio U.: ‘Information systems and transactions architecture’, International Journal of Policy Analysis and Information Systems, vol. 5, no. 4, 1981, pp. 305-324. Ciborra, Claudio U.: ‘Reframing the role of computers in organizations: The transaction costs approach’, in L. Gallegos; R. Welke; and J. Wetherbe (eds.): Proceedings of the Sixth International Conference on Information Systems, 16-18 December 1985, Indianapolis, Indiana, 1985, pp. 57-69. 448 Cooperative Work and Coordinative Practices Ciborra, Claudio U.; and Margrethe H. Olson: ‘Encountering electronic work groups: A Transaction Costs perspective’, in I. Greif and L. A. Suchman (eds.): CSCW’88: Proceedings of the Conference on Computer-Supported Cooperative Work, 26-28 September 1988, Portland, Oregon, ACM Press, New York, 1988, pp. 94-101. Cicourel, Aaron V.: ‘The integration of distributed knowledge in collaborative medical diagnosis’, in J. Galegher; R. E. Kraut; and C. Egido (eds.): Intellectual Teamwork: Social and Technological Foundations of Cooperative Work, Lawrence Erlbaum, Hillsdale, New Jersey, 1990, pp. 221-242. Ciolfi, Luigina; Geraldine Fitzpatrick; and Liam J. Bannon (eds.): ‘Settings for collaboration: The role of place’. [Special theme of] Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 17, no. 2-3, April-June 2008. Coase, Ronald H.: ‘The nature of the firm’, Economica N.S., vol. 4, no. 4, November 1937, pp. 386-405. Cohen, Gerald Allan: Karl Marx’s Theory of History: A Defence, Clarendon Press, Oxford, 1978. Cohen, Jack; and Ian Stewart: The Collapse of Chaos: Discovering Simplicity in a Complex World, Penguin Books, New York, 1994. Cohen, Paul J.: ‘Comments on the foundations of set theory’ (Proceedings of the Symposium in Pure Mathematics of the American Mathematical Society, held at the University of California, Los Angeles, California, July 10-August 5, 1967, 1967). In D. S. Scott (ed.): Axiomatic Set Theory. American Mathematical Society, Providence, Rhode Island, 1971, pp. 9-15. Cole, Arthur H.; and George B. Watts: The Handicrafts of France as Recorded in the Descriptions des Arts et Metiérs, 1761-1788, Baker Library, Harvard Graduate School of Business Administration, Boston, 1952. Coleman, D. C.: ‘Proto-industrialization: A concept too many’, The Economic History Review, New Series, vol. 36, no. 3, August 1984, pp. 435-448. Collier, Bruce: The Little Engines that Could've: The Calculating Machines of Charles Babbage, PhD diss., The Department of History of Science, Harvard University, Cambridge, Mass., August 1970. <http://robroy.dyndns.info/collier/index.html> Committee of General Literature and Education: The Industry of Nations: A Survey of the Existing State of Arts, Machines, and Manufactures. Part II, Society for Promoting Christian Knowledge, London, 1855. Commons, John R.: Institutional Economics, University of Wisconsin Press, Madison, 1934. Condillac, Etienne Bonnot, abbé de: Traité des systêmes, où l'on en démelle les inconvénients et les avantages (La Haye, 1749). Librairie Arthème Fayard, Paris, 1991. Conklin, Jeff; and Michael L. Begeman: ‘gIBIS: A hypertext tool for exploratory policy discussion’, in I. Greif and L. A. Suchman (eds.): CSCW’88: Proceedings of the Conference on Computer-Supported Cooperative Work, 26-28 September 1988, Portland, Oregon, ACM Press, New York, 1988, pp. 140-152. Conklin, Jeff: ‘Design rationale and maintainability’, in B. Shriver (ed.): Proceedings of 22nd Annual Hawaii International Conference on System Sciences, IEEE Computer Society, vol. 2, 1989, pp. 533-539. Conklin, Jeff: ‘Capturing organizational knowledge’, in R. M. Baecker (ed.): Readings in Groupware and Computer-Supported Cooperative Work: Assisting Human-Human Collaboration, Morgan Kaufmann Publishers, San Mateo, Calif., 1993, pp. 561-565. – Originally published in D. Coleman (ed.): Proceedings of Groupware ’92, Morgan Kaufmann, 1992, pp. 133-137. Constant, Edward W., II: The Origins of the Turbojet Revolution, Johns Hopkins University Press, Baltimore, Maryland, 1980. Constant, Edward W., II: ‘Communities and hierarchies: Structure in the practice of science of technology’, in R. Laudan (ed.): The Nature of Technological Knowledge: Are Models of Scientific Change Relevant?, D. Reidel Publishing, Dordrecht, Boston, and London, 1984, pp. 27-46. Cook, Carolyn L.: ‘Streamlining office procedures: An analysis using the information control net model’, in: National Computer Conference, 1980, 1980, pp. 555-565. References 449 Copeland, B. Jack (ed.): The Essential Turing: Seminal Writings in Computing, Logic, Philosophy, Artificial Intelligence, and Artificial Life plus The Secrets of Enigma, Oxford University Press, Oxford, 2004. Copeland, B. Jack (ed.): Alan Turing’s Automatic Computing Engine: The Master Codebreaker’s Struggle to Build the Modern Computer, Oxford University Press, Oxford, 2005. Copeland, B. Jack; and Diane Proudfoot: ‘Turing and the computer’, in B. J. Copeland (ed.): Alan Turing’s Automatic Computing Engine: The Master Codebreaker’s Struggle to Build the Modern Computer, Oxford University Press, Oxford, 2005, pp. 107-148. Copeland, B. Jack (ed.): Colossus: The Secrets of Bletchley Park’s Codebreaking Computers, Oxford University Press, Oxford, 2006a. Copeland, B. Jack: ‘Colossus and the rise of the modern computer’, in B. J. Copeland (ed.): Colossus: The Secrets of Bletchley Park’s Codebreaking Computers, Oxford University Press, Oxford, 2006b, pp. 101-115. Copeland, B. Jack: ‘Introduction’, in B. J. Copeland (ed.): Colossus: The Secrets of Bletchley Park’s Codebreaking Computers, Oxford University Press, Oxford, 2006c, pp. 1-6. Corbató, Fernando J.; Marjorie Merwin-Daggett; and Robert C. Daley: ‘An experimental timesharing system’, in G. A. Barnard (ed.): SJCC’62: Proceedings of the Spring Joint Computer Conference, 1-3 May 1962, San Francisco, California, vol. 21, AFIPS Press, 1962, pp. 335344. Corbató, Fernando J.: ‘Oral history interview’, Interviewed by A. L. Norberg, 18 April 1989, 14 November 1990 1990, at Cambridge, Mass. Interview transcript, Charles Babbage Institute, University of Minnesota, Minneapolis (OH 162). COSMOS: Specification for a Configurable, Structured Message System, Queen Mary College, , London, August 1989. – 68.4 Ext/ALV. Coulter, Jeff: Rethinking Cognitive Theory, Macmillan, London, 1983. Cournot, Antoine Augustin: An Essay on the Foundations of our Knowledge, The Liberal Arts Press, New York, 1956. Transl. from Essai sur les fondements de nos connaissance et sur les caractères de la critique philosophique, Hachette, Paris, 1851. Transl. by M. H. Moore. Crabtree, Andy; Thomas A. Rodden; and Steven D. Benford: ‘Moving with the times: IT research and the boundaries of CSCW’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 14, 2005, pp. 217-251. Crabtree, Andy, et al.: ‘Ethnography considered harmful’, in R. B. Arthur, et al. (eds.): CHI 2009: Proceedings of the 27th international conference on Human factors in computing systems, Boston, Mass., 4-9 April 2009, ACM Press, New York, 2009, pp. 879-888. Crawford, Perry Orson: Automatic Control by Arithmetical Operations, MSc Thesis, Dept. of Physics, Massachusetts Institute of Technology, Cambridge, Mass., 1942. – [The copy of Crawford’s thesis available on-line from MIT is incomplete (it ends on p. 58). According to MIT’s Document Services the ‘thesis file has been corrupted’ and Crawford’s thesis ‘appears to have gone missing from the collection’ (Email communication, 2 and 8 December 2009).]. <http://hdl.handle.net/1721.1/11232> Crawford, Perry Orson: ‘Application of digital computation involving continuous input and output variables’ (5 August 1946). In M. Campbell-Kelly and M. R. Williams (eds.): The Moore School Lectures: Theory and Techniques for Design of Electronic Digital Computers. (Revised ed. of Theory and Techniques for Design of Electronic Computers, 1947-1948). The MIT Press, Cambridge, Mass., 1985, pp. 374-392. Croarken, Mary: ‘Tabulating the heavens: Computing the Nautical Almanac in 18th-century England’, IEEE Annals of the History of Computing, vol. 25, no. 3, July-September 2003, pp. 48-61. Crocker, David H., et al.: ‘RFC 733: Standard for the format of ARPA network text messages’, Request for Comments, no. 733, Network Working Group, 21 November 1977. <http://www.rfc-editor.org/rfc/rfc733.txt> Croft, W. B.; and L. S. Lefkowitz: ‘Task support in an office system’, ACM Transactions on Office Information Systems, vol. 2, 1984, pp. 197-212. Crosby, Alfred W.: The Measure of Reality: Quantification and Western Society, 1250-1600, Cambridge University Press, Cambridge, 1997. 450 Cooperative Work and Coordinative Practices Cummings, Thomas; and Melvin Blumberg: ‘Advanced manufacturing technology and work design’, in T. D. Wall; C. W. Clegg; and N. J. Kemp (eds.): The Human Side of Advanced Manufacturing Technology, John Wiley, Chichester, 1987, pp. 37-60. Cyert, Richard M.; and James G. March: A Behavioral Theory of the Firm, Prentice-Hall, Englewood-Cliffs, New Jersey, 1963. Dahl, Ole-Johan; and Kristen Nygaard: ‘SIMULA: an ALGOL-based simulation language’, Communications of the ACM, vol. 9, no. 9, September 1966, pp. 671-678. Dahrendorf, Ralf: Sozialstruktur des Betriebes, Gabler, Wiesbaden, 1959. Danielsen, Thore, et al.: ‘The AMIGO Project: Advanced group communication model for computer-based communications environment’, in H. Krasner and I. Greif (eds.): CSCW’86: Proceedings. Conference on Computer-Supported Cooperative Work, Austin, Texas, 3-5 December 1986, ACM Press, New York, 1986, pp. 115-142. Darwin, Charles G.: ‘Automatic Computing Engine (ACE)’ (NPL, 17 April 1946). In B. J. Copeland (ed.) Alan Turing’s Automatic Computing Engine: The Master Codebreaker’s Struggle to Build the Modern Computer. Oxford University Press, Oxford, 2005, pp. 53-57. Davenport, Thomas H.: Process Innovation: Reengineering Work through Information Technology, Harvard Business School Press, Boston, Mass., 1993. Davies, Donald Watts: ‘Proposal for the development of a national communication service for online data processing’, Manuscript, [National Physical Laboratory, UK], 15 December 1965. <http://www.cs.utexas.edu/users/chris/DIGITAL_ARCHIVE/NPL/DaviesLetter.html> Davies, Donald Watts: Proposal for a Digital Communication Network, National Physical Laboratory, June 1966. <http://www.cs.utexas.edu/users/chris/DIGITAL_ARCHIVE/NPL/Davies05.pdf> Davies, Donald Watts: ‘Foreword’, in B. J. Copeland (ed.): Alan Turing’s Automatic Computing Engine: The Master Codebreaker’s Struggle to Build the Modern Computer, Oxford University Press, Oxford, 2005, pp. vii-ix. Davis, Philip J.; and Reuben Hersh: The Mathematical Experience, Birkhäuser, Boston, 1980. – Pelican Books, Harmondsworth, England, 1983. De Cindio, Fiorella, et al.: ‘CHAOS as coordination technology’, in H. Krasner and I. Greif (eds.): CSCW’86: Proceedings. Conference on Computer-Supported Cooperative Work, 3-5 December 1986, Austin, Texas, ACM Press, New York, 1986, pp. 325-342. De Cindio, Fiorella, et al.: ‘CHAOS: a knowledge-based system for conversing within offices’, in W.Lamersdorf (ed.): Office Knowledge: Representation, Management and Utilization, Elsevier Science Publishers B.V., North-Holland, Amsterdam, 1988. De Michelis, Giorgio; and M. Antonietta Grasso: ‘How to put cooperative work in context: Analysis and design requirements’, in L. J. Bannon and K. Schmidt (eds.): Issues of Supporting Organizational Context in CSCW Systems, Computing Department, Lancaster University, Lancaster, UK, October 1993, pp. 73-100. COMIC Deliverable D1.1. <ftp://ftp.comp.lancs.ac.uk/pub/comic> de Prony, Gaspard F. C. M. Riche: ‘Notice sur les Grandes Tables Logarithmique &c.’ (Paris, 7 June 1824). In: Recueil des Discourses lus dans la seance publique de L'Academie Royale Des Sciences, 7 Juin 1824. – [Translation entitled ‘On the great logarithmic and trigonometric tables, adapted to the new decimal metric system’ (Science Museum Library, Babbage Archives, Document code: BAB U2l/l). Excerpt probably copied at Charles Babbage’s request.]. <http://sites.google.com/site/babbagedifferenceengine/barondeprony'sdescriptionoftheconstruct i> Degani, Asaf; and Earl L. Wiener: Human Factors of Flight-Deck Checklists: The Normal Checklist, Ames Research Center, National Aeronautics and Space Administration, Moffett Field, California, May 1990. – NASA Contractor Report 177549; Contract NCC2-377. Delayre, Alexandre [Delaire]: ‘Épingle’, in D. Diderot and J. L. R. d’Alembert (eds.): Encyclopédie, ou dictionnaire raisonné des sciences, des arts et des métiers, vol. 5, Briasson, David, Le Breton & Durand, Paris, 1755, pp. 804-807. Text ed. by R. Morrissey. – ARTFL Encyclopédie Projet, University of Chicago, 2008. <http://portail.atilf.fr/encyclopedie/> References 451 Deloney, Thomas: The Pleasant Historie of Iohn Winchcomb, in his Younger Yeares Called Iack of Newbery (London, 1597; 10th edition, 1626). Text ed. by F. O. Mann. In T. Deloney: The Works of Thomas Deloney. Oxford at the Clarendon Press, London etc., 1912, pp. 1-68. Demchak, Chris C.: War, Technological Complexity, and the U.S. Army, PhD diss., Department of Political Science, University of California, Berkeley, Calif., 1987. Descartes, René: Meditations on First Philosophy (Paris, 1641, vol. 2). In R. Descartes: The Philosophical Writings of Descartes. Cambridge University Press, Cambridge, 1984, pp. 1-62. Diderot, Denis: ‘Art’ (Encyclopédie, ou dictionnaire raisonné des sciences, des arts et des métiers, Paris, 1751). In D. Diderot: Œeuvres. Robert Laffont, Paris, 1994, vol. 1 (Philosophie), pp. 265-276. Diebold, John: Automation: The Advent of the Automatic Factory, Van Nostrand, New York, etc., 1952. Divitini, Monica, et al.: A multi-agent approach to the design of coordination mechanisms, Centre for Cognitive Science, Roskilde University, Roskilde, Denmark, 1995. Working Papers in Cognitive Science and HCI. – WPCS-95-5. Divitini, Monica; Carla Simone; and Kjeld Schmidt: ‘ABACO: Coordination mechanisms in a multi-agent perspective’, in: COOP’96: Second International Conference on the Design of Cooperative Systems, 12–14 June 1996, Antibes-Juan-les-Pins, France, INRIA Sophia Antipolis, France, 1996, pp. 103-122. Dourish, Paul; and Victoria Bellotti: ‘Awareness and coordination in shared workspaces’, in M. M. Mantei; R. M. Baecker; and R. E. Kraut (eds.): CSCW’92: Proceedings of the Conference on Computer-Supported Cooperative Work, 31 October–4 November 1992, Toronto, Canada, ACM Press, New York, 1992, pp. 107-114. Dourish, Paul; and Sara Bly: ‘Portholes: supporting awareness in a distributed work group’, in P. Bauersfeld; J. Bennett; and G. Lynch (eds.): CHI’92 Conference Proceedings: ACM Conference on Human Factors in Computing Systems, 3-7 May 1992, Monterey, California, ACM Press, New York, 1992, pp. 541-547. Dourish, Paul, et al.: ‘Freeflow: Mediating between representation and action in workflow systems’, in G. M. Olson; J. S. Olson; and M. S. Ackerman (eds.): CSCW’96: Proceedings of the Conference on Computer-Supported Cooperative Work, 16-20 November 1996, Boston, Mass., ACM Press, New York, 1996, pp. 190-198. Dourish, Paul: ‘Extending awareness beyond synchronous collaboration’, in T. Brinck and S. E. McDaniel (eds.): CHI’97 Workshop on Awareness in Collaborative Systems, 22-27 March 1997, Atlanta, Georgia, 1997. <http://www.usabilityfirst.com/groupware/awareness> Dourish, Paul; and Graham Button: ‘On “technomethodology”: Foundational relationships between ethnomethodology and system design’, Human-Computer Interaction, vol. 13, no. 4, 1998, pp. 395-432. Dourish, Paul: Where the Action is: The Foundation of Embodied Interaction, MIT Press, Cambridge, Mass., 2001. Dourish, Paul: ‘Re-space-ing place: “Place” and “space” ten years on’, in P. J. Hinds and D. Martin (eds.): Proceedings of CSCW'06: The 20th Anniversary Conference on Computer Supported Cooperative Work, Banff, Alberta, Canada, 4-8 November 2006, ACM Press, New York, 2006a, pp. 299-308. Dourish, Paul: ‘Implications for design’, in R. E. Grinter, et al. (eds.): CHI 2006: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 22-27 April 2006, Montréal, Québec, Canada, ACM Press, New York, 2006b, pp. 541-550. Dreyfus, Hubert L.: What Computers Can’t Do: The Limits of Artificial Intelligence, Harper & Row, New York, 1979. (Rev. ed.; 1st ed. 1972). Drury, Maurice O’C.: ‘Conversations with Wittgenstein’ (1974). Text ed. by R. Rhees. In R. Rhees (ed.): Ludwig Wittgenstein: Personal Recollections. Rowman and Littlefield, Totowa, New Jersey, 1981, pp. 112-189. Dubinskas, Frank A.: ‘Cultural constructions: The many faces of time’, in F. A. Dubinskas (ed.): Making Time: Ethnographies of High-Technology Organizations, Temple University Press, Philadelphia, 1988, pp. 3-38. 452 Cooperative Work and Coordinative Practices Eckert, J. Presper: ‘A preview of a digital computing machine’ (15 July 1946). In M. CampbellKelly and M. R. Williams (eds.): The Moore School Lectures: Theory and Techniques for Design of Electronic Digital Computers. (Revised ed. of Theory and Techniques for Design of Electronic Computers, 1947-1948). The MIT Press, Cambridge, Mass., 1985, pp. 109-126. Eckert, Wallace J.: Punched Card Methods in Scientific Computation, The Thomas J. Watson Astronomical Computing Bureau, New York, 1940. – The MIT Press, Cambridge, Mass., and London, 1984. Edgerton, David: The Shock of the Old: Technology and Global History Since 1900, Profile Books, London, 2006. Egger, Edeltraud; and Ina Wagner: ‘Negotiating temporal orders: The case of collaborative time management in a surgery clinic’, Computer Supported Cooperative Work (CSCW): An International Journal, vol. 1, no. 4, 1993, pp. 255-275. Ehn, Pelle: Work-Oriented Design of Computer Artifacts, Arbetslivscentrum, Stockholm, 1988. Ellis, C. A.; S. J. Gibbs; and G. L. Rein: ‘Groupware: Some issues and experiences’, Communications of the ACM, vol. 34, no. 1, January 1991, pp. 38-58. Ellis, Clarence A.: ‘Information control nets: A mathematical model of office information flow’, in: Proceedings of the ACM Conference on Simulation, Measurement and Modeling, August 1979, Boulder, Colorado, ACM, New York, 1979, pp. 225-239. Ellis, Clarence A.; and Gary J. Nutt: ‘Office information systems and computer science’, Computing Surveys, vol. 12, no. 1, March 1980, pp. 27-60. Ellis, Clarence A.; and Marc Bernal: ‘Officetalk-D: An experimental office information system’, in: Proceedings of the ACM SIGOA conference on Office information system, 21-23 June 1982, Philadelphia, Pennsylvannia, 1982, pp. 131-140. Ellis, Clarence A.: ‘Formal and informal models of office activity’, in R. E. A. Mason (ed.): Information Processing ’83. Proceedings of the IFIP 9th World Computer Congress, 19-23 September 1983, Paris, France, North-Holland, Amsterdam, 1983, pp. 11-22. Ellis, Clarence A.; Karim Keddara; and Grzegorz Rozenberg: ‘Dynamic change within workflow systems’, in N. Comstock, et al. (eds.): COOCS’95: Conference on Organizational Computing Systems, 13-16 August 1995, Milpitas, California, ACM Press, New York, 1995, pp. 10-21. Endsley, Mica R.; and Daniel J. Garland (eds.): Situation Awareness Analysis and Measurement, Lawrence Erlbaum Associates, Mahwah, New Jersey, and London, 2000. Endsley, Mica R.; and William M. Jones: ‘A model of inter- and intrateam situation awareness: Implications for design. training, and measurement’, in M. McNeese; E. Salas; and M. R. Endsley (eds.): New Trends in Cooperative Activities: Understanding System Dynamics in Complex Environments, Human Factors and Ergonomics Society, Santa Monica, Calif., 2001, pp. 46-67. Engelbart, Douglas C.: Augmenting human intellect: A conceptual framework (Prepared for: Director of Information Sciences, Air Force Office of Scientific Research, Washington 25, D.C., Contract AF49(638)-1024), Stanford Research Institute, Menlo Park, Calif., October 1962. – AFOSR-3233. <http://www.dougengelbart.org/pubs/augment-3906.html> Engelbart, Douglas C.; and William K. English: ‘A research center for augmenting human intellect’, in: FJCC’68: Proceedings of the Fall Joint Computing Conference, 9-11 December 1968, San Francisco, California. Part I, vol. 33, AFIPS Press, New York, 1968, pp. 395-410. – Also in I. Greif (ed.): Computer-Supported Cooperative Work: A Book of Readings, Morgan Kaufmann Publishers, San Mateo, Calif., 1988, pp. 81-105. Engelbart, Douglas C.: ‘Authorship provisions in AUGMENT’, in: Proc. IEEE Compcon Conference, 1984. – Also in I. Greif (ed.): Computer-Supported Cooperative Work: A Book of Readings, Morgan Kaufmann Publishers, San Mateo, Calif., 1988, pp. 107-126. Engelbart, Douglas C.; and Harvey Lehtman: ‘Working together’, Byte, vol. 13, no. 13, December 1988, pp. 245-252. Engleberger, Joseph E.: Robotics in Practice: Management and Applications of Industrial Robots, Kogan Page, London, 1980. English, W.: ‘The textile industry: Silk production and manufacture, 1750-1900’, in C. Singer, et al. (eds.): A History of Technology, Volume IV: The Industrial Revolution, c 1750 to c 1850, Oxford University Press, Oxford, 1958, pp. 277-307. References 453 English, William K.; Douglas C. Engelbart; and Melwyn L. Berman: ‘Display-selection techniques for text manipulation’, IEEE Transactions on Human Factors in Electronics, vol. HFE-8, no. 1, March 1967, pp. 5-15. Essinger, James: Jacquard’s Web: How a Hand-loom Led to the Birth of the Information Age, Oxford University Press, Oxford, 2005. (Paperback. ed., 2007). Eveland, J. D.; and Tora K. Bikson: ‘Evolving electronic communication networks: an empirical assessment’, Office: Technology and People, vol. 3, no. 2, 1987, pp. 103-128. Exner, Wilhelm Franz: Johann Beckmann: Begründer der technologischen Wissenschaft, Gerold, Wien, 1878. – Kessinger Pub Co, Whitefish, Montana, 2009. – [The reprint title contains a typo: Johann Beckmann: Segründer der technologischen Wissenschaft]. Fayol, Henri: Administration industrielle el générale, Dunod, Paris, 1918, 1999. – Orginally published in Bulletin de la Societé de l’Industrie minérale, 1916. (English translation: General and Industrial Management, Pitman, London, 1949). Farrington, Benjamin: Greek Science, Penguin Books, London, 1944/49. – Spokesman, Nottingham, 2000. Ferry, Georgina: A Computer Called LEO: Lyons Teashops and the World’s First Office Computer, HarperCollins, London, 2003. (Paperback ed., 2004). Fikes, Richard E.; and D. Austin Henderson, Jr.: ‘On supporting the use of procedures in office work’, in: Proceedings of the 1st Annual Conference on Artificial Intelligence, 18-20 August 1980, Stanford University, American Association for Artificial Intelligence, 1980, pp. 202-207. Finholt, Thomas A.; and Stephanie D. Teasley: ‘Psychology: The need for psychology in research on computer-supported cooperative work’, Social Science Computer Review, vol. 16, no. 1, Spring 1998, pp. 40-52. Finley, Moses I.: ‘Was Greek civilization based on slave labour?’ (Historia, 1959). In M. I. Finley (ed.): Slavery in Classical Antiquity: View and Controversies. W. Heffer & Sons, Cambridge, 1960, pp. 53-72. Finley, Moses I.: The Ancient Economy, University of California Press, Berkeley, Calif., 1973. (3rd ed., 1999). Fish, Robert S.; Robert E. Kraut; and Barbara L. Chalfonte: ‘The VideoWindow system in informal communications’, in T. Bikson and F. Halasz (eds.): CSCW’90: Proceedings of the Conference on Computer-Supported Cooperative Work, 7-10 October 1990, Los Angeles, Calif., ACM Press, New York, 1990, pp. 1-11. Fitts, Paul M.: ‘The information capacity of the human motor system in controlling amplitude of movement’, Journal of Experimental Psychology, vol. 47, no. 6, June 1954, pp. 381-391. Fitzpatrick, Geraldine; William J. Tolone; and Simon M. Kaplan: ‘Work, locales and distributed social worlds’, in H. Marmolin; Y. Sundblad; and K. Schmidt (eds.): ECSCW’95: Proceedings of the Fourth European Conference on Computer-Supported Cooperative Work, 10–14 September 1995, Stockholm, Sweden, Kluwer Academic Publishers, Dordrecht, 1995, pp. 1-16. Fitzpatrick, Geraldine; Simon M. Kaplan; and Tim Mansfield: ‘Physical spaces, virtual places and social worlds: A study of work in the virtual’, in G. M. Olson; J. S. Olson; and M. S. Ackerman (eds.): CSCW’96: Proceedings of the Conference on Computer-Supported Cooperative Work, 16-20 November 1996, Boston, Mass., ACM press, New York, 1996, pp. 334-343. Fitzpatrick, Geraldine, et al.: ‘Supporting public availability and accessibility with Elvin: Experiences and reflections’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 11, no. 3-4, 2002. Flach, John M., et al. (eds.): Global Perspectives on the Ecology of Human-Machine Systems, Lawrence Erlbaum, Hillsdale, New Jersey, 1995. Flores, Fernando, et al.: ‘Computer systems and the design of organizational interaction’, ACM Transactions on Office Information Systems, vol. 6, no. 2, April 1988, pp. 153-172. Freiberger, Paul; and Michael Swaine: Fire in the Valley: The Making of the Personal Computer, McGraw-Hill, New York, etc., 2000. (2nd ed.; 1st ed. 1984). Fuchs, Ludwin; and Wolfgang Prinz: ‘Aspects of organizational context in CSCW’, in L. J. Bannon and K. Schmidt (eds.): Issues of Supporting Organizational Context in CSCW Systems, Computing Department, Lancaster University, Lancaster, UK, October 1993, pp. 11-47. COMIC Deliverable D1.1. <ftp://ftp.comp.lancs.ac.uk/pub/comic> 454 Cooperative Work and Coordinative Practices Fuchs, Ludwin; Uta Pankoke-Babatz; and Wolfgang Prinz: ‘Supporting cooperative awareness with local event mechanisms: The GroupDesk system’, in H. Marmolin; Y. Sundblad; and K. Schmidt (eds.): ECSCW’95: Proceedings of the Fourth European Conference on ComputerSupported Cooperative Work, 10–14 September 1995, Stockholm, Sweden, Kluwer Academic Publishers, Dordrecht, 1995, pp. 245-260. Gannon, Paul: Colossus: Bletchley Park’s Greatest Secret, Atlantic Books, London, 2006. (Paperback ed., 2007). Garfinkel, Harold: Studies in Ethnomethodology, Prentice-Hall, Englewood-Cliffs, New Jersey, 1967. – Polity Press, Cambridge, 1987. Garfinkel, Harold; and Egon Bittner: ‘“Good” organizational reasons for “bad” clinic records’ (1967). In H. Garfinkel: Studies in Ethnomethodology. Prentice-Hall, Englewood-Cliffs, New Jersey, 1967, pp. 186-207. Garlan, Yvon: Slavery in Ancient Greece, Cornell University Press, Ithaca and London, 1988. Transl. from Les esclaves en Grèce ancienne, François Maspero, Paris, 1982. Transl. by J. Lloyd. (Revised ed.). Garnsey, Peter: Ideas of Slavery from Aristotle to Augustine, Cambridge University Press, Cambridge, 1996. Gasser, Les; and Alan H. Bond: ‘An analysis of problems and research in distributed artificial intelligence’, in L. Gasser and A. H. Bond (eds.): Readings in Distributed Artificial Intelligence, Morgan Kaufmann Publishers, San Mateo, Calif., 1988, pp. 3-35. Gaver, William W.: ‘Sound support for collaboration’, in L. J. Bannon; M. Robinson; and K. Schmidt (eds.): ECSCW’91: Proceedings of the Second European Conference on ComputerSupported Cooperative Work, 24–27 September 1991, Amsterdam, Kluwer Academic Publishers, Dordrecht, 1991, pp. 293-308. Gaver, William W.: ‘The affordances of media spaces for collaboration’, in M. M. Mantei; R. M. Baecker; and R. E. Kraut (eds.): CSCW’92: Proceedings of the Conference on ComputerSupported Cooperative Work, 31 October–4 November 1992, Toronto, Canada, ACM Press, New York, 1992, pp. 17-24. Gaver, William W., et al.: ‘Realizing a video environment: EuroPARC’s RAVE system’, in P. Bauersfeld; J. Bennett; and G. Lynch (eds.): CHI’92 Conference Proceedings: ACM Conference on Human Factors in Computing Systems, 3-7 May 1992, Monterey, California, ACM Press, New York, 1992, pp. 27-35. Gaver, William W.: ‘Provocative awareness’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 11, no. 3-4, 2002, pp. 475-493. Geertz, Clifford: The Interpretation of Culture: Selected Essays, Basic Books, New York, 1973. Gell, Alfred: The Anthropology of Time: Cultural Constructions of Temporal Maps and Images, Berg, Oxford/Providence, 1992. Gerrard, Steve: ‘Wittgenstein’s philosophies of mathematics’, Synthese, vol. 87, 1991, pp. 125142. Gerson, Elihu M.; and Susan Leigh Star: ‘Analyzing due process in the workplace’, ACM Transactions on Office Information Systems, vol. 4, no. 3, July 1986, pp. 257-270. Gerson, Elihu M.: [Personal communication]. August 1989. Gibson, James J.: The Ecological Approach to Visual Perception, Houghton-Mifflin, Boston, 1979. – Lawrence Erlbaum Associates, Hillsdale, New Jersey, 1986. Giedion, Siegfried: Mechanization Takes Command: A Contribution to Anonymous History, Oxford University Press, New York, 1948. Gilbert, K. R.: ‘Machine-tools’, in C. Singer, et al. (eds.): A History of Technology, Volume IV: The Industrial Revolution, c 1750 to c 1850, Oxford University Press, Oxford, 1958, pp. 417441. Gillespie, Richard: Manufacturing Knowledge: A History of the Hawthorne Experiments, Cambridge University Press, Cambridge, 1991. (Paperback ed., 1993). Gillies, James; and Robert Cailliau: How the Web was Born: The Story of the World Wide Web, Oxford University Press, Oxford, 2000. References 455 Gillispie, Charles Coulston (ed.): A Diderot Pictorial Encyclopedia of Trades and Industry: Manufacturing and the Technical Arts in Plates Selected from ‘L’Encyclopedie, ou dictionnaire raissoné des sciences, des arts et des métiers’ of Denis Diderot, vol. 1-2, Dover Publications, New York, 1959. Transl. from Recueil de planches, sur les sciences, les arts liberaux, et les arts méchaniques, avec leur explication, Paris, 1763-72. Goffman, Erving: Encounters: Two Studies in the Sociology of Interaction, Bobbs-Merrill, Indianapolis, 1961. Goffman, Erving: Behavior in Public Places: Notes on the Social Organization of Gatherings, The Free Press, New York, 1963. Goldstine, Herman H.: The Computer from Pascal to von Neumann, Princeton University Press, Princeton, New Jersey, 1972. Good, Jack; Donald Michie; and Geoffrey Timms: General Report on Tunny: With Emphasis on Statistical Methods, Government Code and Cypher School, 1945. vol. 1-2. – Public Record Office, HW 25/4 & HW 25/4. – Bibliograpical data based on <http://www.alanturing.net/tunny_report/. — Partial transcript by Tony Sale, March 2001: http://www.codesandciphers.org.uk/documents/newman/newman.pdf> Goody, Jack: The Domestication of the Savage Mind, Cambridge University Press, Cambridge, 1977. Goody, Jack: The Logic of Writing and the Organization of Society, Cambridge University Press, Cambridge, 1986. Goody, Jack: The Interface Between the Written and the Oral, Cambridge University Press, Cambridge, 1987. Goody, Jack: The Power of the Written Tradition, Smithsonian Institution Press, Washington D.C., 2000. Grattan-Guinness, Ivor: ‘Work for the hairdressers: The production of de Prony’s logartihmic and trigonometric tables’, IEEE Annals of the History of Computing, vol. 12, no. 3, July-September 1990, pp. 177-185. Greenberg, Saul: ‘Sharing views and interactions with single-user applications’, in: COIS’90: Proceedings of the Conference on Office Information Systems, April 1990, Boston, 1990. Greenberg, Saul: ‘[Introduction to] Computer-supported Cooperative Work and Groupware’, in S. Greenberg (ed.): Computer-supported Cooperative Work and Groupware, Academic Press, London, 1991, pp. 1-8. Greif, Irene; and Sunil K. Sarin: ‘Data sharing in group work’, in H. Krasner and I. Greif (eds.): CSCW’86: Proceedings. Conference on Computer-Supported Cooperative Work, Austin, Texas, 3-5 December 1986, ACM Press, New York, 1986, pp. 175-183. Greif, Irene (ed.): Computer-Supported Cooperative Work: A Book of Readings, Morgan Kaufmann Publishers, San Mateo, Calif., 1988a. Greif, Irene: ‘Overview’, in I. Greif (ed.): Computer-Supported Cooperative Work: A Book of Readings, Morgan Kaufmann Publishers, San Mateo, Calif., 1988b, pp. 5-12. Grier, David Alan: ‘Human computers: The first pioneers of the information age’, Endeavour, vol. 25, no. 1, March 2001, pp. 28–32. Grier, David Alan: When Computers Were Human, Princeton University Press, Princeton and Oxford, 2005. Grinter, Rebecca E.: ‘Supporting articulation work using software configuration management systems’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 5, no. 4, 1996a, pp. 447-465. Grinter, Rebecca E.: Understanding Dependencies: A Study of the Coordination Challenges in Software Development, PhD diss., University of California, Irvine, 1996b. Grønbæk, Kaj; and Preben Mogensen: ‘Informing general CSCW product development through cooperative design in specific work domains’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 6, no. 4, 1997, pp. 275-304. Groover, Mikell P.: Automation, Production Systems, and Computer-Aided Manufacturing, Prentice-Hall, Englewood-Cliffs, New Jersey, 1980. 456 Cooperative Work and Coordinative Practices Gross, Tom; and Wolfgang Prinz: ‘Awareness in context: A light-weight approach’, in K. Kuutti, et al. (eds.): ECSCW 2003: Proceedings of the Eighth European Conference on Computer Supported Cooperative Work, 14–18 September 2003, Helsinki, Finland, Kluwer Academic Publishers, Dordrecht, 2003, pp. 295-314. Gross, Tom; and Wolfgang Prinz: ‘Modelling shared contexts in cooperative environments: Concept, implementation, and evaluation’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 13, no. 3-4, August 2004, pp. 283-303. Gruber, Tom: ‘Ontology’, in L. Liu and M. T. Özsu (eds.): Encyclopedia of Database Systems, Springer-Verlag, 2009. Grudin, Jonathan: ‘Why groupware applications fail: Problems in design and evaluation’, Office: Technology and People, vol. 4, no. 3, 1989, pp. 245-264. Grudin, Jonathan: ‘CSCW: The convergence of two development contexts’, in S. P. Robertson; G. M. Olson; and J. S. Olson (eds.): CHI’91 Conference Proceedings: ACM SIGCHI Conference on Human Factors in Computing Systems, New Orleans, Lousiana, 27 Apri – 2 May 1991, ACM Press, New York, N.Y., 1991, pp. 91-97. Grudin, Jonathan: ‘Computer-Supported Cooperative Work: History and focus’, IEEE Computer, vol. 27, no. 5, May 1994, pp. 19-26. Grudin, Jonathan: ‘CSCW and Groupware: Their history and trajectory’, in Y. Matsushita (ed.): Designing Communication and Collaboration Support Systems, Gordon and Breach Science Publishers, Amsterdam, 1999, pp. 1-15. Grudin, Jonathan: ‘The GUI shock: Computer graphics and human-computer interaction’, Interactions, March-April 2006a, pp. 45-47, 55. Grudin, Jonathan: ‘The demon in the basement’, Interactions, November-December 2006b, pp. 50-53. Gunn, Thomas G.: Manufacturing for Competitive Advantage. Becoming a World Class Manufacturer, Ballinger, Cambridge, Mass., 1987. Gutwin, Carl: Workspace Awareness in Real-Time Distributed Groupware, PhD dissertation, Department of Computer Science, The Universit of Calgary, Calgary, Alberta, December 1997. Gutwin, Carl; and Saul Greenberg: ‘The effects of workspace awareness support on the usability of real-time distributed groupware’, ACM Transactions on Computer-Human Interaction, vol. 6, no. 2, September 1999, pp. 243-281. Gutwin, Carl; and Saul Greenberg: ‘A descriptive framework of workspace awareness for realtime groupware’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 11, no. 3-4, 2002. Haas, Christina: Writing Technology: Studies on the Materiality of Literacy, Lawrence Erlbaum, Mahwah, New Jersey, 1996. Hacker, Peter M. S.: Insight and Illusion: Themes in the Philosophy of Wittgenstein, Thoemmes Press, Bristol, England, 1989. (2nd ed.; 1st ed. 1972). Hacker, Peter M. S.: ‘Malcolm on language and rules’ (Philosophy, 1990). In P. M. S. Hacker: Wittgenstein: Connections and Controversies. Clarendon Press, Oxford, 2001, pp. 310-323. Hafner, Katie; and Matthew Lyon: Where Wizards Stay up Late: The Origins of the Internet, Simon & Schuster, London, etc., 2003. (2nd ed.; 1st ed., 1996). Hammer, Marius: Vergleichende Morphologie der Arbeit in der europäischen Automobilindustrie: Die Entwickling sur Automation, 1959. Hammer, Michael; and Jay S. Kunin: ‘Design principles of an office specification language’, in D. Medley (ed.): AFIPS’80: Proceedings of the National Computer Conference, 19-22 May 1980, Anaheim, California, vol. 49, AFIPS Press, Arlington, Virginia, 1980, pp. 541-547. Hammer, Michael; and Marvin Sirbu: ‘What is office automation?’, in: Proceedings. First Office Automation Conference, 3-5 March 1980, Atlanta, Georgia, 1980. Hardy, Ian R.: The Evolution of ARPANET email, History thesis paper, University of California at Berkeley, Berkeley, Calif., 13 May 1996. <http://www.ifla.org.sg/documents/internet/hari1.txt> References 457 Harper, Richard H. R.; John A. Hughes; and Dan Z. Shapiro: ‘Working in harmony: An examination of computer technology in air traffic control’, in P. Wilson; J. M. Bowers; and S. D. Benford (eds.): ECSCW’89: Proceedings of the 1st European Conference on Computer Supported Cooperative Work, 13-15 September 1989, Gatwick, London, London, 1989a, pp. 73-86. Harper, Richard H. R.; John A. Hughes; and Dan Z. Shapiro: The Functionality of Flight Strips in ATC Work: The report for the Civil Aviation Authority, Lancaster Sociotechnics Group, Department of Sociology, Lancaster University, Lancaster, U.K., January 1989b. Harper, Richard H. R.; John A. Hughes; and Dan Z. Shapiro: ‘Harmonious working and CSCW: Computer technology and air traffic control’, in J. M. Bowers and S. D. Benford (eds.): Studies in Computer Supported Cooperative Work. Theory, Practice and Design, North-Holland, Amsterdam, 1991, pp. 225-234. Harper, Richard H. R.; and John A. Hughes: ‘What a f—ing system! Send ’em all to the same place and then expect us to stop ’em hitting: Managing technology work in air traffic control’, in G. Button (ed.): Technology in Working Order: Studies of Work, Interaction, and Technology, Routledge, London and New York, 1993, pp. 127-144. Harrington, Joseph: Computer Integrated Manufacturing, Krieger, Malabar, Florida, 1979. Harrington, Joseph: Understanding the Manufacturing Process. Key to Successful CAD/CAM Implementation, Marcel Dekker, New York, 1984. Harris, Roy: The Origin of Writing, Duckworth, London, 1986. Harris, Roy: Signs of Writing, Routledge, London and New York, 1995. Harris, Roy: Rethinking Writing, Indiana University Press, Bloomington and Indianapolis, 2000. Harrison, Steve; and Paul Dourish: ‘Re-placing space: The roles of place and space in collaborative systems’, in G. M. Olson; J. S. Olson; and M. S. Ackerman (eds.): CSCW’96: Proceedings of the Conference on Computer-Supported Cooperative Work, 16-20 November 1996, Boston, Mass., ACM Press, New York, 1996, pp. 67-76. Hauben, Michael; and Ronda Hauben: Netizens: On the History and Impact of Usenet and the Internet, IEEE Computer Society Press, Los Alamitos, Calif., 1997. Heart, Frank, et al.: Draft ARPANET Completion Report, [Bound computer printout, unpublished], Bolt Beranek and Newman, 9 September 1977. – Quoted from Abbate: Inventing the Internet, The MIT Press, Cambridge, Mass., and London, 1999. Heart, Frank, et al.: ARPANET Completion Report, Bolt Beranek and Newman, 4 January 1978. (Also published as A History of the ARPANET: The First Decade, BBN 1981, Report #4799.). <http://www.cs.utexas.edu/users/chris/DIGITAL_ARCHIVE/ARPANET/DARPA4799.pdf> Heath, Christian C.; and Paul Luff: ‘Collaborative activity and technological design: Task coordination in London Underground control rooms’, in L. J. Bannon; M. Robinson; and K. Schmidt (eds.): ECSCW’91: Proceedings of the Second European Conference on ComputerSupported Cooperative Work, 24–27 September 1991, Amsterdam, Kluwer Academic Publishers, Dordrecht, 1991a, pp. 65-80. Heath, Christian C.; and Paul Luff: ‘Disembodied conduct: Communication through video in a multi-media office environment’, in S. P. Robertson; G. M. Olson; and J. S. Olson (eds.): CHI’91 Conference Proceedings: ACM SIGCHI Conference on Human Factors in Computing Systems, New Orleans, Lousiana, 27 April-2 May 1991, ACM Press, New York, N.Y., 1991b, pp. 99-103. Heath, Christian C.; and Paul Luff: ‘Collaboration and control: Crisis management and multimedia technology in London Underground control rooms’, Computer Supported Cooperative Work (CSCW): An International Journal, vol. 1, no. 1-2, 1992a, pp. 69-94. Heath, Christian C.; and Paul Luff: ‘Media space and communicative asymmetries: Preliminary observations of video mediated interaction’, Human-Computer Interaction, vol. 7, 1992b, pp. 315-346. Heath, Christian C.; and Paul Luff: ‘Disembodied conduct: Interactional asymmetries in videomediated communication’, in G. Button (ed.): Technology in Working Order: Studies of Work, Interaction, and Technology, Routledge, London and New York, 1993, pp. 35-54. Heath, Christian C., et al.: ‘Unpacking collaboration: the interactional organisation of trading in a City dealing room’, Computer Supported Cooperative Work (CSCW): An International Journal, vol. 3, no. 2, 1995a, pp. 147-165. 458 Cooperative Work and Coordinative Practices Heath, Christian C.; Paul Luff; and Abigail J. Sellen: ‘Reconsidering the virtual workplace: Flexible support for collaborative activity’, in H. Marmolin; Y. Sundblad; and K. Schmidt (eds.): ECSCW’95: Proceedings of the Fourth European Conference on Computer-Supported Cooperative Work, 10–14 September 1995, Stockholm, Sweden, Kluwer Academic Publishers, Dordrecht, 1995b, pp. 83-99. Heath, Christian C.; and Paul Luff: ‘Convergent activities: Line control and passenger information on the London Underground’, in Y. Engeström and D. Middleton (eds.): Cognition and Communication at Work, Cambridge University Press, Cambridge, 1996, pp. 96-129. Heath, Christian C., et al.: ‘Configuring awareness’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 11, no. 3-4, 2002. Henderson, Kathryn: On Line and On Paper: Visual representations, Visual Culture, and Computer Graphics in Design Engineering, MIT Press, Cambridge, Mass., and London, 1999. Heritage, John C.: Garfinkel and Ethnomethodology, Polity Press, Cambridge, 1984. Hertzfeld, Andy: Revolution in the Valley, O’Reilly, Sebastopol, Calif., 2005. Hewitt, Carl: ‘Viewing control structures as patterns of passing messages’, Artificial Intelligence, vol. 8, 1977, pp. 323-364. Hewitt, Carl: ‘Offices are open systems’, ACM Transactions on Office Information Systems, vol. 4, no. 3, July 1986, pp. 271-287. Hilbert, David: ‘Mathematical problems’ (International Congress of Mathematicians, Paris, 1900). Transl. from ‘Mathematische Probleme’. Transl. by M. F. Winston. In J. J. Gray: The Hilbert Challenge. Oxford University Press, Oxford, 2000, pp. 240-282. – [German original in Archiv für Mathematik und Physik, vol. 1, 1901; Enlish transl. in Bulletin of the American Mathematical Society, vol. 8, 1902] Hilbert, David: ‘On the infinite’ (Münster, 4 June 1925). Transl. from ‘Über das Unendliche’. Transl. by S. Bauer-Mengelberg. In J. van Heijenoort (ed.) From Frege to Gödel: A Source Book in Mathematical Logic, 1879–1931. Harvard University Press, Cambridge, Mass., 1967, pp. 367-392. Hilbert, David: ‘The foundations of mathematics’ (Hamburg, July 1927). Transl. by S. BauerMengelberg and D. Føllesdal. In J. van Heijenoort (ed.) From Frege to Gödel: A Source Book in Mathematical Logic, 1879–1931. Harvard University Press, Cambridge, Mass., 1967, pp. 464-479. Hillis, W. Daniel: The Pattern on the Stone: The Simple Idea that Make Computers Work, Weidenfeld & Nicolson, London, 1998. – Phoenix, London, 2001 (Paperback ed.). Hiltz, Starr Roxanne; and Murray Turoff: The Network Nation: Human Communication via Computer, Addison-Wesley, Reading, Mass., 1978. – The MIT Press, Cambridge, Mass., 1993 (Rev. ed.). Hiltz, Starr Roxanne; and Murray Turoff: ‘The evolution of user behavior in a computerized conferencing system’, Communications of the ACM, vol. 24, no. 11, November 1981, pp. 739751. Hiltz, Starr Roxanne; and Elaine B. Kerr: Studies of Computer Mediated Communications Systems: A Synthesis of the Findings: Final Report on a Workshop Sponsored by the Division of Information Science and Technology, National Science Foundation, Computerized Conferencing and Communications Center, New Jersey Institute of Technology, Newark, New Jersey, 1981. – Research Report Number 16. <http://library.njit.edu/archives/cccc-materials/> Hiltz, Starr Roxanne; and Murray Turoff: ‘Structuring computer-mediated communication systems to avoid information overload’, Communications of the ACM, vol. 28, no. 7, July 1985, pp. 680-689. Hirschhorn, Larry: Beyond Mechanization: Work and Technology in a Postindustrial Age, MIT Press, Cambridge, Mass., and London, 1984. Hobbes, Thomas: Leviathan, or The Matter, Forme, & Power of a Common-Wealth Ecclesiaticall and Civill, London, 1651. – Penguin Books, Harmondsworth, Middlesex, 1968. Hodgskin, Thomas: Labour Defended Against the Claims of Capital, Or, The Unproductiveness of Capital Proved with Reference to the Present Combinations Amongst Journeymen, Knight & Lacey, London, 1825. – Augustus M. Kelley, Publishers, New York, 1969. Holland, John H.: Hidden Order: How Adaption Builds Complexity, Addison-Wesley Publishing Co., Reading, Mass., 1995. References 459 Holt, Anatol W.: ‘Coordination Technology and Petri Nets’, in G. Rozenberg (ed.): Advances in Petri Nets 1985, vol. 222, Springer-Verlag, Berlin, 1985, pp. 278-296. Text ed. by G. Goos and J. Hartmanis. Howard, Robert: ‘Systems design and social responsibility: The political implications of ‘Computer-Supported Cooperative Work’: A commentary’, Office: Technology and People, vol. 3, no. 2, 1987. Howse, Derek: Greenwich Time and the Discovery of the Longitude, Oxford University Press, Oxford, 1980. – Philip Wilson Publishers, London, 1997. Hughes, John A., et al.: The Automation of Air Traffic Control, Lancaster Sociotechnics Group, Department of Sociology, Lancaster University, Lancaster, UK, October 1988. Hughes, John A.; David W. Randall; and Dan Z. Shapiro: ‘CSCW: Discipline or paradigm? A sociological perspective’, in L. J. Bannon; M. Robinson; and K. Schmidt (eds.): ECSCW’91: Proceedings of the Second European Conference on Computer-Supported Cooperative Work, 24–27 September 1991, Amsterdam, Kluwer Academic Publishers, Dordrecht, 1991, pp. 309323. Hughes, Thomas P.: ‘The evolution of large technological systems’, in W. E. Bijker; T. P. Hughes; and T. J. Pinch (eds.): The Social Construction of Technological Systems: New Directions in the Sociology and History of Technology, The MIT Press, Cambridge, Mass., and London, 1987, pp. 51-82. (Paperback ed., 1989). Hull, David L.: Science as a Process: An Evolutionary Account of the Social and Conceptual Development of Science, The University of Chicago Press, Chicago, 1988. Hutchins, Edwin L.; James D. Hollan; and Donald A. Norman: ‘Direct manipulation interfaces’, in D. A. Norman and S. W. Draper (eds.): User Centered System Design, Lawrence Erlbaum, New Jersey, 1986, pp. 87-124. Hutchins, Edwin L.: ‘Mediation and automatization’, Quarterly Newsletter of the Laboratory of Comparative Human Cognition [University of California, San Diego], vol. 8, no. 2, April 1986, pp. 47-58. Hutchins, Edwin L.: ‘The social organization of distributed cognition’, in L. B. Resnick; J. M. Levine; and S. D. Teasley (eds.): Perspectives on Socially Shared Cognition, American Psychological Association, Washington, DC, 1991, pp. 283-307. Hutchins, Edwin L.: Cognition in the Wild, The MIT Press, Cambridge, Mass., and London, England, 1995. Hutchins, Edwin L.; and Tove Klausen: ‘Distributed cognition in an airline cockpit’, in Y. Engeström and D. Middleton (eds.): Cognition and Communication at Work, Cambridge University Press, Cambridge, 1996, pp. 15-34. Ishii, Hiroshi: ‘TeamWorkStation: Towards a seamless shared workspace’, in T. Bikson and F. Halasz (eds.): CSCW’90: Proceedings of the Conference on Computer-Supported Cooperative Work, 7-10 October 1990, Los Angeles, Calif., ACM Press, New York, 1990, pp. 13-26. Ishii, Hiroshi; Minoru Kobayaski; and Jonathan Grudin: ‘Integration of inter-personal space and shared workspace: ClearBoard design and experiments’, in M. M. Mantei; R. M. Baecker; and R. E. Kraut (eds.): CSCW’92: Proceedings of the Conference on Computer-Supported Cooperative Work, 31 October–4 November 1992, Toronto, Canada, ACM Press, New York, 1992, pp. 33-42. Jirotka, Marina, et al. (eds.): ‘Special Issue: Collaboration in e-Research’. [Special theme of] Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 15, no. 4, August 2006. Johannsen, Gunnar, et al.: ‘Final report of experimental psychology group’, in N. Moray (ed.): Mental Workload: Its Theory and Measurement, Plenum Press, New York and London, 1979, pp. 101-114. Johansen, Robert: Groupware: Computer Support for Business Teams, Free Press, New York and London, 1988. Johnson, Allen W.; and Timothy Earle: The Evolution of Human Societies: From Foraging Group to Agrarian State, Stanford University Press, Stanford, Calif., 1987. Johnson, Jeff, et al.: ‘The Xerox Star: A Retrospective’, IEEE Computer, vol. 22, no. 9, September 1989, pp. 11-26, 28-29. 460 Cooperative Work and Coordinative Practices Johnson, Philip: ‘Supporting exploratory CSCW with the EGRET framework’, in M. M. Mantei; R. M. Baecker; and R. E. Kraut (eds.): CSCW’92: Proceedings of the Conference on Computer-Supported Cooperative Work, 31 October–4 November 1992, Toronto, Canada, ACM Press, New York, 1992, pp. 298-305. Johnson-Lenz, Peter; and Trudy Johnson-Lenz: ‘Groupware: the process and impacts of design choices’, in E. B. Kerr and S. R. Hiltz (eds.): Computer-Mediated Communication Systems, Academic Press, New York, 1982. Johnston, William A.; and Veronica J. Dark: ‘Selective attention’, Annual Review of Psychology, vol. 37, 1986, pp. 43-75. Jones, Stephen R. G.: ‘Was there a Hawthorne effect?’, The American Journal of Sociology, vol. 98, no. 3, November 1992, pp. 451-468. Jordan, Nehemiah: ‘Allocation of functions between man and machines in automated systems’, Journal of Applied Psychology, vol. 47, no. 3, June 1963, pp. 161-165. Kaavé, Bjarne: Undersøgelse af brugersamspil i system til produktionsstyring, M.Sc. diss., Technical University of Denmark, Lyngby, Denmark, 1990. Kanigel, Robert: The One Best Way: Frederick Winslow Taylor and the Enigma of Efficiency, The MIT Press, Cambridge, Mass., and London, 1997. (Paperback ed., 2005). Kant, Immanuel: ‘Über den Gemeinspruch: Das mag in der Theorie richtig sein, taugt aber nicht für die Praxis’ (Berlinische Monatsschrift, September 1793). Text ed. by W. Weischedel. In I. Kant: Werke in zwölf Bänden. Suhrkamp Verlag, Frankfurt a. M., 1964, vol. XI, pp. 125-172. Kantowitz, Barry H.; and Robert D. Sorkin: ‘Allocation of functions’, in G. Salvendy (ed.): Handbook of Human Factors, John Wiley, New York etc., 1987, pp. 355-369. Kaplan, Simon M., et al.: ‘Flexible, active support for collaborative work with Conversation Builder’, in M. M. Mantei; R. M. Baecker; and R. E. Kraut (eds.): CSCW’92: Proceedings of the Conference on Computer-Supported Cooperative Work, 31 October–4 November 1992, Toronto, Canada, ACM Press, New York, 1992, pp. 378-385. Kaptelinin, Victor: ‘Computer-mediated activity: Functional organs in social and developmental contexts’, in B. A. Nardi (ed.): Context and Consciousness: Activity Theory and HumanComputer Interaction, The MIT Press, Cambridge, Mass., 1997, pp. 45-68. Kasbi, Catherine; and Maurice de Montmollin: ‘Activity without decision and responsibility: The case of nuclear power plants’, in J. Rasmussen; B. Brehmer; and J. Leplat (eds.): Distributed Decision Making. Cognitive Models for Cooperative Work, John Wiley & Sons, Chichester, 1991, pp. 275-283. Kauffman, Stuart: At Home in the Universe: The Search for Laws of Self-Organization and Complexity, Oxford University Press, New York and Oxford, 1995. Kawell, Jr., Leonard , et al.: ‘Replicated document management in a group communication system’, in I. Greif and L. A. Suchman (eds.): CSCW’88: Proceedings of the Conference on Computer-Supported Cooperative Work, 26-28 September 1988, Portland, Oregon, ACM Press, New York, 1988, pp. 395[-404]. Kelley, Charles R.: Manual and Automatic Control: A Theory of Manual Control and its Application to Manual and to Automatic Systems, Wiley, New York, 1968. Kern, Horst; and Michael Schumann: Industriearbeit und Arbeiterbewußtsein: Eine empirische Untersuchung über den Einfluß der aktuellen technischen Entwicklung auf die industrielle Arbeit und das Arbeiterbewußtsein, vol. 1-2, Europäische Verlagsanstalt, Frankfurt am Main, 1970. Kerr, Elaine B.; and Starr Roxanne Hiltz: Computer-Mediated Communication Systems: Status and Evaluation, Academic Press, Orlando, etc., 1982. Kiesler, Sara; Jane Siegel; and Timothy W. McGuire: ‘Social psychological aspects of computermediated communication’, American Psychologist, vol. 39, no. 10, October 1984, pp. 11231134. Kling, Rob: ‘Social analyses of computing: Theoretical perspectives in recent empirical research’, Computing Surveys, vol. 12, no. 1, March 1980, pp. 61-110. Kling, Rob: ‘Cooperation, coordination and control in computer-supported work’, Communications of the ACM, vol. 34, no. 12, December 1991, pp. 83-88. Kochhar, A. K.; and N. D. Burns: Microprocessors and their Manufacturing Applications, Edward Arnold, London, 1983. References 461 Koren, Yoram: Computer Control of Manufacturing Systems, McGraw-Hill, New York, etc., 1983. Korsch, Karl: Karl Marx, Chapman & Hall, London, 1938. – Russell & Russell, New York, 1963. Kraut, Robert E.; Carmen Egido; and Jolene Galegher: ‘Patterns of contact and communication in scientific research collaboration’, in J. Galegher; R. E. Kraut; and C. Egido (eds.): Intellectual Teamwork: Social and Technological Foundations of Cooperative Work, Lawrence Erlbaum, Hillsdale, New Jersey, 1990, pp. 149-171. Kreifelts, Thomas: ‘DOMINO: Ein System zur Abwicklung arbeitsteiliger Vorgänge im Büro’, Angewandte Informatik, vol. 26, no. 4, 1984, pp. 137-146. Kreifelts, Thomas, et al.: ‘Experiences with the DOMINO office procedure system’, in L. J. Bannon; M. Robinson; and K. Schmidt (eds.): ECSCW’91: Proceedings of the Second European Conference on Computer-Supported Cooperative Work, 24–27 September 1991, Amsterdam, Kluwer Academic Publishers, Dordrecht, 1991a, pp. 117-130. Kreifelts, Thomas, et al.: ‘A design tools for autonomous agents’, in J. M. Bowers and S. D. Benford (eds.): Studies in Computer Supported Cooperative Work. Theory, Practice and Design, North-Holland, Amsterdam, 1991b, pp. 131-144. Kreifelts, Thomas; Elke Hinrichs; and Gerd Woetzel: ‘Sharing to-do lists with a distributed task manager’, in G. De Michelis; C. Simone; and K. Schmidt (eds.): ECSCW’93: Proceedings of the Third European Conference on Computer-Supported Cooperative Work, 13-17 September 1993, Milano, Italy, Kluwer Academic Publishers, Dordrecht, 1993, pp. 31-46. Kuhn, Thomas S.: The Structure of Scientific Revolutions (1962; 2nd ed. 1969). University of Chicago Press, Chicago, 1969. Kuutti, Kari: ‘Activity Theory as a potential framework for Human-Computer Interaction research ’, in B. A. Nardi (ed.): Context and Consciousness: Activity Theory and HumanComputer Interaction, The MIT Press, Cambridge, Mass., 1997, pp. 17-44. La Porte, Todd R. (ed.): Organized Social Complexity: Challenge to Politics and Policy, Princeton University Press, Princeton, New Jersey, 1975. La Porte, Todd R.; and Paula M. Consolini: ‘Working in practice but not in theory: Theoretical challenges of “high-reliability organizations”’, Journal of Public Administration Research and Theory, vol. 1, no. 1, 1991, pp. 19-47. Laiserin, Jerry: ‘The convergence of CAD standards’. In Digital Architect, 2002. Resource April. – Reprint, Reprint City, Reprint 2002. Lakatos, Imre: ‘A renaissance of empiricism in the recent philosophy of mathematics?’ (1967). Text ed. by J. Worrall and G. Currie. In I. Lakatos: Mathematics, Science and Epistemology: Philosophical Papers, Volume 2. (Paperback ed., 1980). Cambridge University Press, Cambridge, 1978, pp. 24-42. Lakatos, Imre: ‘Falsification and the methodology of scientific research programmes’, in I. Lakatos and A. Musgrave (eds.): Criticism and the Growth of Knowledge: Proceedings of the International Colloquium in the Philosophy of Science, London, 1965, Cambridge University Press, Cambridge, 1970, pp. 91-196. – Also in Imre Lakatos: The Methodology of Scientific Research Programmes: Philosophical Papers, Volume 1, Cambridge University Press. Cambridge, 1978, pp. 8-101. Lampson, Butler W.: ‘Personal distributed computing: The Alto and Ethernet software’, in A. Goldberg (ed.): A History of Personal Workstations, Addison-Wesley, Reading, Mass., 1988, pp. 293-335. Landes, David S.: Revolution in Time: Clocks and the Making of the Modern World, Harvard University Press, Cambridge, Mass., 1983. Latour, Bruno: Science in Action: How to follow scientists and engineers through society, Harvard University Press, Cambridge, Mass., 1987. Laudan, Rachel: ‘Introduction’, in R. Laudan (ed.): The Nature of Technological Knowledge: Are Models of Scientific Change Relevant?, D. Reidel Publishing, Dordrecht, Boston, and London, 1984, pp. 1-26. 462 Cooperative Work and Coordinative Practices Lauwers, J. Chris; and Keith A. Lantz: ‘Collaboration awareness in support of collaboration transparency: Requirements for the next generation of shared window systems’, in J. C. Chew and J. Whiteside (eds.): CHI’90 Conference Proceedings: ACM SIGCHI Conference on Human Factors in Computing Systems, Seattle, Washington, 1-5 April 1990, ACM Press, New York, N.Y., 1990, pp. 303-311. Lave, Jean: ‘Situated learning in communities of practice’, in L. B. Resnick; J. M. Levine; and S. D. Teasley (eds.): Perspectives on socially shared cognition, Americal Psychological Association, Washington, DC, 1991, pp. 63-82. Law, John; and John Hassard (eds.): Actor Network Theory and After, Blackwell, London, 1999. Layton, Edwin T.: ‘Technology as knowledge’, Technology and Culture, vol. 15, no. 1, January 1974, pp. 31-41. LeCuyer, A.: ‘Design on the Computer: Frank O. Gehry und Peter Eisenman’, ARCH+, no. 128, September 1995, pp. 26-29. Lee, Charlotte P., et al. (eds.): ‘Special Issue: Scientific collaboration through cyberinfrastructure’. [Special theme of] Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 19, 2010. – [Forthcoming]. Lévi-Strauss, Claude: La pensée sauvage, Librairie Plon, Paris, 1962. – Plon, Paris, 1990. Licklider, Joseph Carl Robnett: ‘Man-computer symbiosis’ (IRE Transactions on Human Factors in Electronics, March 1960). In R. W. Taylor (ed.): In Memoriam: J. C. R. Licklider, 19151990. Digital Systems Research Center, Palo Alto, Calif., 1990, pp. 4–11. <[Also in Adele Goldberg: A History of Personal Workstations, Addison-Wesley, Reading, Mass., 1988, pp. 131-140]> Licklider, Joseph Carl Robnett; and Welden E. Clark: ‘On-line man-computer communication’, in G. A. Barnard (ed.): SJCC’62: Proceedings of the Spring Joint Computer Conference, 1-3 May 1962, San Francisco, California, vol. 21, AFIPS Press, 1962, pp. 113-128. Lindgren, Michael: Glory and Failure: The Difference Engines of Johann Müller, Charles Babbabe and Georg and Edvard Scheutz, Linköping University, Linköping, 1987. List, Friedrich: Das nationale System der politischen Ökonomie, Cotta Verlag, Stuttgart u. Tübingen, 1841a. List, Friedrich: The National System of Political Economy, J. B. Lippincott, Philadelphia, 1856. Transl. from Das nationale System der politischen Ökonomie, Cotta Verlag, Stuttgart u. Tübingen, 1841b. Transl. by G. A. Matile. Lojek, Bo: History of Semiconductor Engineering, Springer, Berlin-Heidelberg, 2007. Luff, Paul; Jon Hindmarsh; and Christian C. Heath (eds.): Workplace Studies: Recovering Work Practices and Informing System Design, Cambridge University Press, Cambridge, 2000. Luke, Hugh D.: Automation for Productivity, John Wiley & Sons, New York etc., 1972. Lutters, Wayne G.; and Mark S. Ackerman: ‘Achieving safety: A field study of boundary objects in aircraft technical support’, in E. Churchill, et al. (eds.): CSCW 2002: ACM Conference on Computer Supported Cooperative Work, 16 - 20 November 2002, New Orleans, Louisiana, ACM Press, New York, 2002, pp. 266-275. Lynch, Michael: ‘Extending Wittgenstein: The Pivotal Move from Epistemology to the Sociology of Science’, in A. Pickering (ed.): Science as Practice and Culture, Chicago University Press, Chicago, 1993, pp. 215-265. Maas, W.; and J. Van Rijs: FRANAX: Excursions on Density, O10 Publishers, Rotterdam, 1998. Mahoney, Michael S.: ‘Computers and mathematics: The search for a discipline of computer science’, in J. Echeverria; A. Ibarra; and T. Mormann (eds.): The Space of Mathematics: Philosophical, Epistemological, and Historical Explorations, Walter de Gruyter, Berlin and New York, 1992, pp. 349-363. Mahoney, Michael S.: ‘The histories of computing(s)’, Interdisciplinary Science Reviews, vol. 30, no. 2, 2005, pp. 119-135. Malcolm, Norman: Nothing is Hidden: Wittgenstein’s Criticism of Hs Early Thought, Basil Blackwell, Oxford, 1986. Malcolm, Norman: Wittgensteinian Themes: Essays, 1978-1989, Cornell University Press, Ithaca and London, 1995. Text ed. by G. H. von Wright. Malone, Thomas W.; JoAnne Yates; and Robert I. Benjamin: ‘Electronic markets and electronic hierarchies’, Communications of the ACM, vol. 30, no. 6, June 1987, pp. 484-497. References 463 Malone, Thomas W.; and Kevin Crowston: ‘What is coordination theory and how can it help design cooperative work systems’, in T. Bikson and F. Halasz (eds.): CSCW’90: Proceedings of the Conference on Computer-Supported Cooperative Work, 7-10 October 1990, Los Angeles, Calif., ACM Press, New York, 1990, pp. 357-370. Malone, Thomas W.; Hum-Yew Lai; and Christopher Fry: ‘Experiments with Oval: A radically tailorable tool for cooperative work’, in M. M. Mantei; R. M. Baecker; and R. E. Kraut (eds.): CSCW’92: Proceedings of the Conference on Computer-Supported Cooperative Work, 31 October–4 November 1992, Toronto, Canada, ACM Press, New York, 1992, pp. 289-297. Malone, Thomas W.; and Kevin Crowston: The Interdisciplinary Study of Coordination, Center of Coordination Science, MIT, Cambridge, Mass., 1992. Malone, Thomas W., et al.: ‘Tools for inventing organizations: Toward a handbook of organizational processes’, in: Proceedings of the 2nd IEEE Workshop on Enabling Technologies Infrastructure for Collaborative Enterprises, 20-22 April 1993, Morgantown, West Virginia, 1993. Malone, Thomas W.; Hum-Yew Lai; and Christopher Fry: ‘Experiments with Oval: A radically tailorable tool for cooperative work’, ACM Transactions on Office Information Systems, vol. 13, no. 2, April 1995, pp. 177-205. Mantei, Marilyn M., et al.: ‘Experiences in the use of a media space’, in S. P. Robertson; G. M. Olson; and J. S. Olson (eds.): CHI’91 Conference Proceedings: ACM SIGCHI Conference on Human Factors in Computing Systems, New Orleans, Lousiana, 27 April-2 May 1991, ACM Press, New York, N.Y., 1991, pp. 203-208. March, James G.; and Herbert A. Simon: Organizations, John Wiley, New York, 1958. Mariani, John A.; and Wolfgang Prinz: ‘From multi-user to shared object systems: Awareness about co-workers in cooperation support object databases’, in H. Reichel (ed.): GI-Tagung: Informatik - Wirtschaft - Gesellschaft, Springer, Berlin-Heidelberg, 1993, pp. 476-481. <http://www.fit.fraunhofer.de/~prinz/papers/gi93-awareness.html> Mark, Gloria: ‘Conventions and commitments in distributed CSCW groups’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 11, no. 3-4, 2002. Marx, Karl: ‘Thesen über Feuerbach’ (Manuscript, Spring 1845a). In K. Marx and F. Engels: Werke. Dietz Verlag, Berlin, 1962, pp. 5-7. Marx, Karl: Brüsseler Hefte 1845 (Manuscript, 1845b). Text ed. by G. A. Bagaturija, et al. In K. Marx and F. Engels: Gesamtausgabe (MEGA➁). Akademie Verlag, Berlin, 1998, vol. IV/3, pp. 111-433. Marx, Karl: Misère de la philosophie: Reponse a La philosophie de la misère de M. Proudhon, A. Frank / C. G. Vogler, Paris / Bruxelles, 1847. Marx, Karl: ‘Der achtzehnte Brumaire des Louis Bonaparte’ (Die Revolution: Eine Zeitscrhift in zwanglosen Heften, New York, 1852). Text ed. by M. Hundt, et al. In K. Marx and F. Engels: Gesamtausgabe (MEGA➁). Dietz Verlag, Berlin, 1985, vol. I/11, pp. 96-189. Marx, Karl: ‘Einleitung’ (Manuscript, August 1857). Text ed. by V. K. Brušlinskij; L. R. Mis’kevič; and A. G. Syrov. In K. Marx and F. Engels: Gesamtausgabe (MEGA➁). Dietz Verlag, Berlin, 1976-81, vol. II/1.1 (Ökonomische Manuskripte 1857/58), pp. 17-45. – English transl. by Nicolaus, in Marx: Grundrisse, Pelican, Harmondsworth, 1973. Marx, Karl: Grundrisse der Kritik der politischen Ökonomie (Manuscript, 1857-58a). Text ed. by V. K. Brušlinskij; L. R. Mis’kevič; and A. G. Syrov. In K. Marx and F. Engels: Gesamtausgabe (MEGA➁). Dietz Verlag, Berlin, 1976-1981, vol. II/1, pp. 47-747. Marx, Karl: Grundrisse der Kritik der politischen Ökonomie (Rohentwurf), 1857-1858. Anhang 1850-1859 (Manuscript, 1857-58b). Dietz Verlag, Berlin, 1953. Marx, Karl: Zur Kritik der politischen Ökonomie (Manuskript 1861-1863) (1861-63). Text ed. by A. Schnickmann, et al. In K. Marx and F. Engels: Gesamtausgabe (MEGA➁). Dietz Verlag, Berlin, 1976-1982, vol. II/3.1-II/3.6. Marx, Karl: [Letter to Friedrich Engels], London, 28 January 1863. In K. Marx and F. Engels: Werke. Dietz Verlag, Berlin, 1964, vol. 30, pp. 319-323. Marx, Karl: ‘Ökonomische Manuscripte 1863-1867’ (1863-67). In K. Marx and F. Engels (eds.): Gesamtausgabe (MEGA). 464 Cooperative Work and Coordinative Practices Marx, Karl: Das Kapital. Kritik der politischen Ökonomie. Erster Band. Buch I: Der Produktionsprocess des Kapitals (Hamburg, 1867a). Text ed. by E. Kopf, et al. In K. Marx and F. Engels: Gesamtausgabe (MEGA➁). Dietz Verlag, Berlin, 1983, vol. II/5. Marx, Karl: Das Kapital. Kritik der politischen Ökonomie. Erster Band. Buch I: Der Produktionsprocess des Kapitals (Hamburg, 1867b; 4th ed., 1890). Text ed. by F. Engels. In K. Marx and F. Engels: Werke. Dietz Verlag, Berlin, 1962, vol. 23. Marx, Karl: Capital: A Critique of Political Economy. Book One: The Process of Production of Capital (Hamburg, 1867c; 4th ed., 1890). Transl. by S. Moore and E. Aveling. Text ed. by F. Engels. In K. Marx and F. Engels: Collected Works. Lawrence & Wishart, London, 1996, vol. 35. Marx, Karl: ‘À la rédaction de l’Otečestvennyâ Zapiski’ (Draft letter, in French, OctoberNovember 1877). Text ed. by H. Schwab, et al. In K. Marx and F. Engels: Gesamtausgabe (MEGA➁). Dietz Verlag, Berlin, 1985, vol. I/25, pp. 112-117. Marx, Karl: ‘Lettre à Vera Ivanovna Zasulič (Projet I-IV)’ (Draft letters, in French, 18 February 8 March 1881). Text ed. by H. Schwab, et al. In K. Marx and F. Engels: Gesamtausgabe (MEGA➁). Dietz Verlag, Berlin, 1985, vol. I/25, pp. 217-240. Mayr, Ernst; and Peter D. Ashlock: Principles of Systematic Zoology, McGraw-Hill, New York, 1991. (2nd ed.; 1st ed. 1969). Mayr, Ernst; and Walter J. Bock: ‘Classifications and other ordering systems’, Journal of Zoological Systematics & Evolutionary Research, vol. 40, no. 4, December 2002, pp. 169-194. Mayr, Otto: The Origins of Feedback Control, The MIT Press, Cambridge, Mass., and London, 1970. Transl. from Zur Frühgeschichte der technischen Regelungen, R. Oldenbourg Verlag, Wien, 1969. McCarthy, John: ‘Reminiscences on the history of time sharing’, Stanford University, Winter or Spring 1983. <http://www-formal.stanford.edu/jmc/history/timesharing/timesharing.html> McKenney, James L.: Waves of Change: Business Evolution through Information Technology, Harvard Business School Press, Boston, Mass., 1995. McNeese, Michael; Eduardo Salas; and Mica R. Endsley (eds.): New Trends in Cooperative Activities: Understanding System Dynamics in Complex Environments, Human Factors and Ergonomics Society, Santa Monica, Calif., 2001. McPherson, J. C.: ‘Mathematical operations with punched cards’, Journal of the American Statistical Association, vol. 37, no. 218, June 1942, pp. 275-281. Medina-Mora, Raul, et al.: ‘The Action Workflow approach to workflow management technology’, in M. M. Mantei; R. M. Baecker; and R. E. Kraut (eds.): CSCW’92: Proceedings of the Conference on Computer-Supported Cooperative Work, 31 October–4 November 1992, Toronto, Canada, ACM Press, New York, 1992, pp. 281-288. Mendels, Franklin F.: ‘Proto-industrialization: The first phase of the industrialization process’, The Journal of Economic History, vol. 32, no. 1, March 1972, pp. 241-261. Merchant, M. Eugene: ‘Flexible Manufacturing Systems: Robotics and computerized automation’, The Annals of the American Academy of Political and Social Science, vol. 470, no. 1, 1983, pp. 123-135. Mickler, Otfried; Eckhard Dittrich; and Uwe Neumann: Technik, Arbeitsorganisation und Arbeit: Eine empirische Untersuchung in der automatischen Produktion, Aspekte Verlag, Frankfurt am Main, 1976. Mill, John Stuart: Principles of Political Economy, with Some of Their Applications to Social Philosophy, John W. Parker, London, 1848. Miller, George Armitage: ‘What is information measurement?’, The American Psychologist, vol. 8, January 1953, pp. 3-11. Miller, George Armitage: ‘The magical number seven, plus or minus two: Some limits on our capacity for processing information’, Psychological Review, vol. 63, no. 2, March 1956, pp. 81-97. Miller, George Armitage; Eugene Galanter; and Karl H. Pribram: Plans and the Structure of Behavior, Holt, Rinehart & Winston, New York, 1960. Mills, Charles Wright: White Collar: The American Middle Classes, Oxford University Press, New York, 1951. References 465 Mindell, David A.: Between Human and Machine: Feedback, Control, and Computing before Cybernetics, Johns Hopkins University Press, Baltimore and London, 2002. Mintzberg, Henry: The Structuring of Organizations: A Synthesis of the Research, Prentice-Hall, Englewood-Cliffs, New Jersey, 1979. Mitchell, W. J.. Picture Theory: Essays on Verbal and Visual Representation, The University of Chicago Press, Chicago, 1994. Mitrofanov, Sergei Petrovich: Scientific Principles of Group Technology, National Lending Library for Science and Technology, Boston Spa, Yorkshire, England. Transl. from Nauchnye osnovy gruppovoi tekhnologii, Leningrad, 1959, 1959. Transl. by E. Harris. Text ed. by T. J. Grayson. 1966. Monden, Yasuhiro: Toyota Production System: Practical Approach to Production Management, Industrial Engineering and Management Press, Institute of Industrial Engineers, Norcross, Georgia, 1983. Monk, Ray: Bertrand Russell: The Spirit of Solitude, Jonathan Cape, London, 1996. – Vintage, 1997 (Paperback ed.). Monod, Jacques: Chance and Necessity: An Essay on the Natural Philosophy of Modern Biology, Collins, London, 1972. Transl. from Le hasard et la nécessité. Essai sur la philosophie naturelle de la biologie moderne, Editions Seuil, Paris, 1970. Transl. by A. Wainhouse. Moore Jr., Barrington: Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making of the Modern World, Boston, 1966. – Penguin University Books, Harmondsworth, 1973. Moray, Neville (ed.): Mental Workload: Its Theory and Measurement, Plenum Press, New York and London, 1979a. Moray, Neville: ‘Models and measures of mental workload’, in N. Moray (ed.): Mental Workload: Its Theory and Measurement, Plenum Press, New York and London, 1979b, pp. 13-21. Morse, Philip M.; and George E. Kimball: Methods of Operations Research, The Technology Press of MIT, Cambridge, Mass., 1951. (rev.ed.; 1st ed., US Navy, 1946). – Dover Publications, Mineola, New York, 2003. Multhauf, Robert P.: ‘Some observations on the state of the history of technology’, Technology and Culture, vol. 15, no. 1, January 1974, pp. 1-12. Mumford, Lewis: Technics and Civilization, Harcourt Brace & Co., New York, 1934. – Harcourt Brace & Co., San Diego, etc., 1963 (Paperback ed.). Myer, Theodore H.; and David Dodds: ‘Notes on the development of message technology’, in D. M. Austin (ed.): Berkeley Workshop on Distributed Data Management and Computer Networks, 1976, Lawrence Berkeley Laboratories, 1976, pp. 144-154. – LBL-5315. Napper, R. B. E.: ‘The Manchester Mark I Computers’, in R. Rojas and U. Hashagen (eds.): The First Computers: History and Architectures, The MIT Press, Cambridge, Mass., and London, 2000, pp. 365-378. Nardi, Bonnie A.: A Small Matter of Programming: Perspectives on End User Computing, The MIT Press, Cambridge, Mass., 1993. Nardi, Bonnie A.; Steve Whittaker; and Erin Bradner: ‘Interaction and outeraction: Instant messaging in action’, in W. A. Kellogg and S. Whittaker (eds.): CSCW 2000: ACM Conference on Computer Supported Cooperative Work, 2–6 December 2000, Philadelphia, Pennsylvania, USA, ACM Press, New York, 2000, pp. 79-88. Neisser, Ulric: Cognition and Reality: Principles and Implications of Cognitive Psychology, W. H. Freeman & Co., San Francisco, 1976. Newton, Janice: ‘Technology and cooperative labour among the Orokaiva’, Mankind, vol. 15, no. 3, December 1985, pp. 214-222. Nicolis, Grégoire; and Ilya Prigogine: Exploring Complexity: An Introduction, W. H. Freeman & Co., New York, 1989. Norberg, Arthur L.: ‘High-technology calculation in the early 20th century: Punched card machinery in business and government’, Technology and Culture, vol. 31, no. 4, October 1990, pp. 753-779. Norman, Donald A.; and Edwin L. Hutchins: Computation via direct manipulation, Institute for Cognitive Science, University of California, San Diego, La Jolla, California, 1 August 1988. – ONR Contract N00014-85-C-0133. 466 Cooperative Work and Coordinative Practices Norman, Donald A.: ‘Cognitive artifacts’, in J. M. Carroll (ed.): Designing Interaction: Psychology at the Human-Computer Interface, Cambridge University Press, Cambridge, 1991, pp. 17-38. Nuland, Sherwin B.: [Review of] A. Gawande: Complications: A Surgeon’s Notes on an Imperfect Science (Metropolitan Books), The New York Review of Books, vol. 49, no. 12, 18 July 2002, pp. 10-13. Nurminen, Markku I.: People or Computers: Three ways of looking at information systems, Studentlitteratur, Lund, Sweden, 1988. Nuttgens, Patrick: The Story of Architecture, Phaidon Press, London, 1997. (2nd ed.; 1st ed. 1983). Nygaard, Kristen; and Ole-Johan Dahl: ‘The development of the SIMULA languages’, ACM SIGPLAN Notices, vol. 13, no. 8, August 1978, pp. 245-272. O’Neill, Judy Elizabeth: The Evolution of Interactive Computing through Time-sharing and Networking, Ph.D. dissertation, University of Minnesota, 1992. Odgaard, Irene: “Praksis var værre end forventet”: Producktionsgrupper på Ålestrup: Grundfos, Denmark. Casestudie-rapport om nye former for arbejdsorganisation, IMF Project on new forms of work organization. Unpublished report, Specialarbejderforbundet i Danmark (SiD), København, 26 August 1994. Odgaard, Irene, et al.: Produktionsgrupper: organisationsudvikling og IT-støtte, CO-industri etc., København, 1999. Ohmae, Kenichi: Triad Power: The Coming Shape of Global Competition, Free Press, New York, 1985. Ohno, Taiichi: Toyota Production System: Beyond Large-scale Production, Productivity Press, New York, 1988. Olson, David R.: The World on Paper: The Conceptual and Cognitive Implications of Writing and Reading, Cambridge University Press, Cambridge, 1994. Olson, Gary M.; and Judith S. Olson: ‘Groupware and computer-supported cooperative work’, in J. A. Jacko and A. Sears (eds.): The Human-Computer Interaction Handbook: Fundamentals, Evolving Technologies and Emerging Applications, Lawrence Erlbaum, Mahwah, New Jersey, 2003, pp. 583-595. Opitz, H.; W. Eversheim; and H. P. Wiendahl: ‘Workpiece classification and its industrial application’, International Journal of Machine Tool Design Research, vol. 9, 1969, pp. 39-50. Opitz, H.: A Classification System to Describe Workpieces, Pergamon Press, Oxford, 1970. Transl. from Werkstückbeschreibendes Klassifizierungssystem, Verlag W. Girardet, Essen. Transl. by R. A. A. Taylor. Text ed. by W. R. MacConnell. Orbiteam: ‘BSCW at work: References’, August 2009. <http://www.bscw.de/english/references.html> Orlikowski, Wanda J.: ‘Learning from NOTES: Organizational issues in groupware implementation’, in M. M. Mantei; R. M. Baecker; and R. E. Kraut (eds.): CSCW’92: Proceedings of the Conference on Computer-Supported Cooperative Work, 31 October–4 November 1992, Toronto, Canada, ACM Press, New York, 1992, pp. 362-369. Orton, John: Semiconductors and the Information Revolution: Magic Crystals that Made IT Happen, Academic Press, Amsterdam, etc., 2009. Ouchi, William G.: ‘Markets, bureaucracies, and clans’, Administrative Science Quarterly, vol. 25, March 1980, pp. 129-141. Palme, Jacob: ‘You have 134 unread mail! Do you want to read them now?’, in H. T. Smith (ed.): Proceedings of the IFIP Conference on Computer Based Message Services, Nottingham, U.K., Elsevier North-Holland, New York, 1984, pp. 175-184. Panko, Raymond R.: ‘The outlook for computer mail’, Telecommunications Policy, June 1977, pp. 242-253. Panko, Raymond R.: ‘Electronic mail’, in J. Slonim; E. A. Unger; and P. S. Fisher (eds.): Advances in Data Communications Management, vol. 2, John Wiley & Sons, New York, 1984, pp. 205-222. Pankoke-Babatz, Uta (ed.): Computer Based Group Communication: The AMIGO Activity Model, Ellis Horwood Publishers, Chichester, 1989a. References 467 Pankoke-Babatz, Uta: ‘Preface’, in U. Pankoke-Babatz (ed.): Computer Based Group Communication: The AMIGO Activity Model, Ellis Horwood Publishers, Chichester, 1989b, pp. 13-16. Pankoke-Babatz, Uta: ‘Group communication and electronic communication’, in U. PankokeBabatz (ed.): Computer Based Group Communication: The AMIGO Activity Model, Ellis Horwood Publishers, Chichester, 1989c, pp. 17-66. Peaucelle, Jean-Louis: ‘From Taylorism to post-Taylorism: Simultaneously pursuing several management objectives’, Journal of Organizational Change Management, vol. 13, no. 5, 2000, pp. 452-467. Peaucelle, Jean-Louis: ‘Adam Smith’s use of multiple references for his pin making example’, The European Journal of the History of Economic Thought, vol. 13, no. 4, 2006, pp. 489-512. Peaucelle, Jean-Louis; and Stéphane Manin: ‘Billettes and the economic viability of pin-making in 1700’, in: Eleventh World Congress of Accounting Historians, July 2006, Nantes, 2006. <http://adamsmithslostlegacy.com/ftp.adamsmithslostlegacy.com/Pinmaking-Billettes.pdf> Peaucelle, Jean-Louis: Adam Smith et la division du travail: La naissance d'une fausse idée, L’Harmattan, Paris, 2007. Perronet, Jean-Rodolphe: ‘Epinglier: Description de la façon dont on fabrique les épingles à Laigle en Normandie’, in D. Diderot and J. L. R. d’Alembert (eds.): Recueil de planches, sur les sciences, les arts liberaux, et les arts méchaniques, avec leur explication.Troisième livraison, 298 Planches, Briasson, David & Le Breton, Paris, 1765, pp. 5:1-5:8. Text ed. by R. Morrissey. – ARTFL Encyclopédie Projet, University of Chicago, 2008. – [Encyclopédie…, vol. 21]. <<http://portail.atilf.fr/encyclopedie/>> Perrow, Charles: ‘The Organizational Context of Human Factors Engineering’, Administrative Science Quaterly, vol. 28, December 1983, pp. 521-541. Perrow, Charles: Normal Accidents: Living with High-Risk Technologies, Basic Books, New York, 1984. Perrow, Charles: Complex Organizations. A Critical Essay, Random House, New York, 1986. (Third ed.). Petty, William: Another Essay in Political Arithmetick, Concerning the Growth of the City of London, with the Measures, Periods, Causes, and Consequences thereof (1683, vol. II). Text ed. by C. H. Hull. In W. Petty: The Economic Writings of Sir William Petty. Cambridge University Press, Cambridge, 1899, pp. 449-487. – Reprinted by Augustus M. Kelley, New York, 1963. Piel, Gerard, et al. (eds.): Automatic Control, G. Bell and Sons, London, 1955. (2nd ed.; 1st. ed. 1955). Pietroforte, Roberto: ‘Communication and governance in the building process’, Construction Managment and Economics, vol. 15, 1997, pp. 71-82. Pitkin, Hanna Fenichel: Wittgenstein and Justice: On the Significance of Ludwig Wittgenstein for Social and Political Thought, University of California Press, Berkeley, Calif., 1972. (Paperback ed., 1993). Plowman, Lydia; Yvonne Rogers; and Magnus Ramage: ‘What are workplace studies for?’, in H. Marmolin; Y. Sundblad; and K. Schmidt (eds.): ECSCW’95: Proceedings of the Fourth European Conference on Computer-Supported Cooperative Work, 10–14 September 1995, Stockholm, Sweden, Kluwer Academic Publishers, Dordrecht, 1995, pp. 309-324. Pollock, Friedrich: Automation: Materialien zur Beurteilung der ökonomischen and sozialen Folgen, Europäische Verlagsanstalt, Frankfurt a. M., 1964. Popitz, Heinrich, et al.: Technik und Industriearbeit: Soziologische Untersuchungen in der Hüttenindustrie, J. C. B. Mohr, Tübingen, 1957. Poppe, Johann Heinrich Moritz: Geschichte der Technologie seit der Wiederherstellung der Wissenschaften bis an das Ende des achtzehnten Jahrhundert: Erster Band, Röwer, Göttingen, 1807. – Georg Olms Verlag, Hildesheim etc., 1999. Postel, Jonathan Bruce: ‘RFC 808: Summary of computer mail services meeting held at BBN on 10 January 1979’, Request for Comments, no. 808, Network Working Group, 1 March 1982. <http://www.rfc-editor.org/rfc/rfc808.txt> 468 Cooperative Work and Coordinative Practices Potts, Colin; and Lara Catledge: ‘Collaborative conceptual design: A large software project case study’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 5, no. 4, 1996, pp. 415-445. Pougès, Claude, et al.: ‘Conception de collecticiels pour l’aide à la prise de décision en situation d’urgence: La nécessité d’une approche pluridisciplinaire et intégrée’, in B. Pavard (ed.): Systèmes coopératifs: De la modélisation à la conception, Octares Éditions, Toulouse, 1994, pp. 351-375. Powell, Walter W.: ‘Neither market nor hierarchy: Network forms of organization’, in B. M. Staw and L. L. Cummings (eds.): Research in Organizational Behavior, vol. 12, JAI Press, Greenwich, Conn., 1989, pp. 295-336. Prigogine, Ilya; and Isabelle Stengers: La nouvelle alliance: Metamorphose de la science, Paris, 1979. Prinz, Wolfgang: ‘TOSCA: Providing organisational information to CSCW applications’, in G. De Michelis; C. Simone; and K. Schmidt (eds.): ECSCW’93: Proceedings of the Third European Conference on Computer-Supported Cooperative Work, 13-17 September 1993, Milano, Italy, Kluwer Academic Publishers, Dordrecht, 1993, pp. 139-154. Prinz, Wolfgang: ‘NESSIE: An awareness environment for cooperative settings’, in S. Bødker; M. Kyng; and K. Schmidt (eds.): ECSCW’99: Proceedings of the Sixth European Conference on Computer-Supported Cooperative Work, 12–16 September 1999, Copenhagen, Kluwer Academic Publishers, Dordrecht, 1999, pp. 391-410. Prinz, Wolfgang: ‘Industry perspective on Collaborative Environments’, in I. L. Ballesteros (ed.): New Collaborative Working Environments 2020: Report on Industry-led FP7 Consultations and 3rd Report of the Experts Group on Collaboration@Work, European Commisssion, DG Information Society and Media, Bruxelles, February 2006, pp. 5-22. Pycock, James; and Wes W. Sharrock: ‘The fault report form: Mechanisms of interaction in design and development project work’, in K. Schmidt (ed.): Social Mechanisms of Interaction, Computing Department, Lancaster University, Lancaster, UK, September 1994, pp. 257-294. COMIC Deliverable D3.2. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Pycock, James: ‘Mechanisms of interaction and technologies of representation: Examining a case study’, in K. Schmidt (ed.): Social Mechanisms of Interaction, Computing Department, Lancaster University, Lancaster, UK, September 1994, pp. 123-148. COMIC Deliverable D3.2. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Quarterman, John S.: The Matrix: Computer Networks and Conferencing Systems Worldwide, Digital Press, Bedford, Mass., 1990. Randall, David W.; Richard H. R. Harper; and Mark Rouncefield: Fieldwork for Design: Theory and Practice, Springer, London, 2007a. Randall, David W., et al.: ‘Ontology building as practical work: Lessons from CSCW’, in: Proceedings of the third International Conference on e-Social Science, 7-9 October 2007, Ann Arbor, Michigan, 2007b. Rasmussen, Jens: On the Structure of Knowledge: A Morphology of Mental Models in a ManMachine Context, Risø National Laboratory, Roskilde, Denmark, November 1979. Rasmussen, Jens; and Morten Lind: Coping with complexity, Risø National Laboratory, Roskilde, Denmark, June 1981. – Risø-M-2293, DK-4000 Roskilde, Denmark. Rasmussen, Jens: ‘The role of hierarchical knowledge representation in decisionmaking and system management’, IEEE Transactions on Systems, Man, and Cybernetics, vol. SMC-15, no. 2, March/April 1985, pp. 234-243. Rasmussen, Jens: ‘A cognitive engineering approach to the modelling of decision making and its organization in process control, emergency management, CAD/CAM, office systems, library systems’, in W. B. Rouse (ed.): Advances in Man-Machine Systems Research, vol. 4, JAI Press, Greenwich, Conn., 1988, pp. 165-243. Rasmussen, Jens; Berndt Brehmer; and Jacques Leplat (eds.): Distributed Decision Making: Cognitive Models for Cooperative Work, John Wiley & Sons, Chichester, 1991. Réaumur, René Antoine Ferchault de; Henri Louis Duhamel du Monceau; and Jean-Rodolphe Perronet: Art de l'épinglier par M. de Reaumur; avec des additions de M. Duhamel du Monceau, & des remarques extraites des memoires de M. Perronet, inspecteur general des ponts & chaussees, Saillant et Noyon, Paris, 1761. References 469 Redlow, Götz: Theoria: Theoretische und praktische Lebensauffassung im philosophischen Denken der Antike, VEB Deutscher Verlag der Wissenschaften, Berlin, 1966. Redmond, Kent C.; and Thomas M. Smith: Project Whirlwind: The History of a Pioneer Computer, Digital Press, Bedford, Mass., 1980. Redmond, Kent C.; and Thomas M. Smith: From Whirlwind to MITRE: The R&D Story of the SAGE Airdefense Computer, The MIT Press, Cambridge, Mass., and London, 2000. Reekie, Fraser: Reekie’s Architectural Drawing, Architectural Press, Amsterdam etc., 1995. Text ed. by T. McCarthy. (4th ed.; 1st ed. 1946). Reid, T. R.: The Chip: How Two Americans Invented the Microchip and Launched a Revolution, Random House, New York, 2001. (2nd ed.; 1st ed., 1985). Resnick, Lauren B.; John M. Levine; and Stephanie D. Teasley (eds.): Perspectives on Socially Shared Cognition, American Psychological Association, Washington D.C., 1991. Reynolds, Peter C.: On the Evolution of Human Behavior: The Argument from Animals to Man, University of California Press, Berkeley, Calif., 1981. Rice, Ronald E.: ‘Computer-mediated communication and organizational innovation’, Journal of Communication, vol. 37, no. 4, Autumn 1987, pp. 65-94. Rice, Ronald E.: ‘Computer-mediated communication system network data: Theoretical concerns and emprirical examples’, International Journal of Man-Machine Studies, vol. 32, no. 6, June 1990, pp. 627-647. Riordan, Michael; and Lillian Hoddeson: Crystal Fire: The Invention of the Transistor and the Birth of the Information Age, W. W. Norton & Co., New York and London, 1997. (Paperback ed., 1998). Rittel, Horst W. J.; and Melvin M. Webber: ‘Dilemmas in a general theory of planning’, Policy Sciences, vol. 4, 1973, pp. 155-169. Roberts, Lawrence G.: ‘Multiple computer networks and intercomputer communication’, in J. Gosden and B. Randell (eds.): SOSP’67: Proceedings of the First ACM Symposium on Operating System Principles, 1-4 October 1967, Gatlinburg, Tennessee, ACM Press, New York, 1967, pp. 3.1-3.6. Roberts, Lawrence G.; and Barry D. Wessler: ‘Computer network development to achieve resource sharing’, in H. L. Cooke (ed.): SJCC’70: Proceedings of the Spring Joint Computer Conference, 5-7 May 1970, Atlantic City, New Jersey, vol. 36, AFIPS Press, Montvale, New Jersey, 1970, pp. 543-549. Robertson, Toni: Designing Over Distance: A Study of Cooperative Work, Embodied Cognition and Technology to Enable remote Collaboration, Submitted for the Degree of Doctor of Philosophy, School of Computing Sciences, University of Technology, Sydney, 1997. 195, pp. Robertson, Toni: ‘The public availability of actions and artefacts’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 11, no. 3-4, 2002. Robinson, Mike: ‘Double-level languages and co-operative working’, AI & Society, vol. 5, 1991, pp. 34-60. Robinson, Mike; and Liam J. Bannon: ‘Questioning representations’, in L. J. Bannon; M. Robinson; and K. Schmidt (eds.): ECSCW’91: Proceedings of the Second European Conference on Computer-Supported Cooperative Work, 24–27 September 1991, Amsterdam, Kluwer Academic Publishers, Dordrecht, 1991, pp. 219-233. Rochlin, Gene I.; Todd R. La Porte; and Karlene H. Roberts: ‘The self-designing high-reliability organization: Aircraft carrier flight operations at sea’, Naval War College Review, Autumn 1987, pp. 76-90. Rochlin, Gene I.: ‘Informal organizational networking as a crisis-avoidance strategy: U. S. naval flight operations as a case study’, Industrial Crisis Quarterly, vol. 3, 1989, pp. 159-176. Rodden, Tom A.; and Gordon Blair: ‘CSCW and distributed systems: The problem of control’, in L. J. Bannon; M. Robinson; and K. Schmidt (eds.): ECSCW’91: Proceedings of the Second European Conference on Computer-Supported Cooperative Work, 24–27 September 1991, Amsterdam, Kluwer Academic Publishers, Dordrecht, 1991, pp. 49-64. Rodden, Tom A.; John A. Mariani; and Gordon Blair: ‘Supporting cooperative applications’, Computer Supported Cooperative Work (CSCW): An International Journal, vol. 1, no. 1-2, 1992, pp. 41-68. 470 Cooperative Work and Coordinative Practices Rodden, Tom A.: ‘Populating the application: A model of awareness for cooperative applications’, in G. M. Olson; J. S. Olson; and M. S. Ackerman (eds.): CSCW’96: Proceedings of the Conference on Computer-Supported Cooperative Work, 16-20 November 1996, Boston, Mass., ACM Press, New York, 1996, pp. 87-96. Roethlisberger, Fritz J.; and William J. Dickson: Management and the Worker, Harvard University Press, Cambridge, Mass., 1939. Rogers, Yvonne: ‘Coordinating computer-mediated work’, Computer Supported Cooperative Work (CSCW): An International Journal, vol. 1, no. 4, 1993, pp. 295-315. Rønby Pedersen, Elin; and Tomas Sokoler: ‘AROMA: Abstract representation of presence supporting mutual awareness’, in S. Pemberton (ed.): CHI’97 Conference Proceedings: ACM SIGCHI Conference on Human Factors in Computing Systems, Atlanta, Georgia, 22-27 March 1997, ACM Press, New York, 1997, pp. 51-58. Roseman, Mark; and Saul Greenberg: ‘TeamRooms: Network places for collaboration’, in G. M. Olson; J. S. Olson; and M. S. Ackerman (eds.): CSCW’96: Proceedings of the Conference on Computer-Supported Cooperative Work, 16-20 November 1996, Boston, Mass., ACM press, New York, 1996, pp. 325-333. Roth, Emilie M.; and David D. Woods: ‘Cognitive task analysis: An approach to knowledge acquisition for intelligent system design’, in G. Guida and C. Tasso (eds.): Topics in Expert System Design. Methodologies and Tools, North-Holland, Amsterdam, 1989, pp. 233-264. Rouse, William B.; and Sandra H. Rouse: ‘Measures of complexity of fault diagnosis tasks’, IEEE Transactions on Systems, Man, and Cybernetics, vol. 9, no. 11, November 1979, pp. 720-727. Russell, Bertrand: ‘The study of mathematics’ (1902). In B. Russell: Mysticism and Logic and Other Essays. George Allen & Unwin, London, 1917, pp. 47-57. – [First published in New Quarterly, November, 1907]. Russell, Bertrand: The Principles of Mathematics, 1903. (2nd ed., 1937). Ryle, Gilbert: The Concept of Mind, Hutchinson’s University Library, London, 1949. Ryle, Gilbert: ‘The thinking of thoughts: What is “le Penseur” doing?’ (University Lectures, University of Saskatchewan, 1968). In G. Ryle: Collected Papers. Volume II: Collected Essays, 1929-1968. Hutchinson & Co, London, 1971, pp. 480-496. Sabbagh, Karl: Skyscraper: The Making of a Building, Macmillan, London, 1989. – Penguin Books, New York, 1991. Salus, Peter H.: Casting the Net: From ARPANET to Internet and Beyond, Addison-Wesley, Reading, Mass., etc., 1995. Sandor, Ovidiu; Cristian Bogdan; and John M. Bowers: ‘Aether: An awareness engine for CSCW’, in J. A. Hughes, et al. (eds.): ECSCW’97: Proceedings of the Fifth European Conference on Computer-Supported Cooperative Work, 7–11 September 1997, Lancaster, U.K., Kluwer Academic Publishers, Dordrecht, 1997, pp. 221-236. Sartre, Jean-Paul: Critique de la raison dialectique. Tome I, Gallimard, Paris, 1960. Savage, Charles M.: Fifth Generation Management for Fifth Generation Technology (A Round Table Discussion), Society of Manufacturing Engineers, Dearborn, Michigan, 1987. Schaffer, Simon: ‘Babbage's calculating engines and the factory system’, Réseaux, vol. 4, no. 2, 1996, pp. 271-298. Schäl, Thomas: ‘System design for cooperative work in the Language Action Perspective: A case study of The Coordinator’, in D. Z. Shapiro; M. Tauber; and R. Traünmuller (eds.): The Design of Computer Supported Cooperative Work and Groupware Systems, North-Holland Elsevier, Amsterdam, 1996, pp. 377-400. Schick, Kathy D.; and Nicholas Toth: Making Silent Stones Speak: Human Evolution and the Dawn of Technology, Simon & Schuster, New York, 1993. Schmidt, Kjeld: ‘Fremmedgørelse og frigørelse’ (1970). In K. Marx: Kritik af den politiske økonomi: Grundrids. Rhodos, København, 1970, pp. 7-19. Schmidt, Kjeld; Kay Clausen; and Erling C. Havn: Computer-integrerede produktionssystemer (CIM), teknologien og dens samfundsmæssige betingelser og konsekvenser, Roskilde University, Roskilde, Denmark, September 1984. – CIM-projektet, working paper, no. 1. <http://cscw.dk/schmidt/papers/cim1984.pdf> References 471 Schmidt, Kjeld: ‘A dialectical approach to functional analysis of office work’, in: IEEE International Conference on Systems, Man, and Cybernetics, 14-17 October 1986, Atlanta, Georgia, IEEE Press, New York, 1986, pp. 1586-1591. Schmidt, Kjeld: Teknikerens arbejde og teknikerens arbejdsplads, Dansk Datamatik Center, Lyngby, Denmark, December 1987. TIA projektet. <http://cscw.dk/schmidt/papers/tia.pdf> Schmidt, Kjeld: ‘Functional Analysis Instrument’, in G. Schäfer, et al. (eds.): Functional Analysis of Office Requirements. A Multiperspective Approach, Wiley, Chichester, 1988a, pp. 261-289. Schmidt, Kjeld: ‘Forms of cooperative work’, in H.-J. Bullinger, et al. (eds.): EURINFO’88: Proceedings of the First European Conference on Information Technology for Organisational Systems, 16-20 May 1988, Athens, Greece, Elsevier Science Publishers (North Holland), 1988b. Schmidt, Kjeld: ‘Cooperative work: A conceptual framework’, in: New Technology, Distributed Decision Making, and Responsibility, 5-7 May 1988, Bad Homburg, Germany, 1988c. Schmidt, Kjeld: Analysis of Cooperative Work. A Conceptual Framework, Risø National Laboratory, Roskilde, Denmark, June 1990. – Risø-M-2890. Schmidt, Kjeld: ‘Riding a tiger, or Computer Supported Cooperative Work’, in L. J. Bannon; M. Robinson; and K. Schmidt (eds.): ECSCW’91: Proceedings of the Second European Conference on Computer-Supported Cooperative Work, 24–27 September 1991, Amsterdam, Kluwer Academic Publishers, Dordrecht, 1991a, pp. 1-16. Schmidt, Kjeld: ‘Computer support for cooperative work in advanced manufacturing’, International Journal of Human Factors in Manufacturing, vol. 1, no. 4, October 1991b, pp. 303-320. Schmidt, Kjeld: ‘Cooperative work: A conceptual framework’, in J. Rasmussen; B. Brehmer; and J. Leplat (eds.): Distributed Decision Making: Cognitive Models for Cooperative Work, John Wiley & Sons, Chichester, 1991c, pp. 75-109. Schmidt, Kjeld; and Liam J. Bannon: ‘CSCW, Or What’s In A Name?’. [Unpublished manuscript]. August 1991. Schmidt, Kjeld; and Liam J. Bannon: ‘Taking CSCW seriously: Supporting articulation work’, Computer Supported Cooperative Work (CSCW): An International Journal, vol. 1, no. 1-2, 1992, pp. 7-40. Schmidt, Kjeld, et al.: ‘Computational mechanisms of interaction: Notations and facilities’, in C. Simone and K. Schmidt (eds.): Computational Mechanisms of Interaction for CSCW, Computing Department, Lancaster University, Lancaster, UK, October 1993, pp. 109-164. COMIC Deliverable D3.1. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Schmidt, Kjeld: ‘Cooperative work and its articulation: Requirements for computer support’, Travail Humain, vol. 57, no. 4, December 1994a, pp. 345-366. Schmidt, Kjeld: Modes and Mechanisms of Interaction in Cooperative Work, Risø National Laboratory, Roskilde, Denmark, 1994b. – Risø-R-666(EN). Schmidt, Kjeld: ‘The organization of cooperative work: Beyond the “Leviathan” conception of the organization of cooperative work’, in J. B. Smith; F. D. Smith; and T. W. Malone (eds.): CSCW’94: Proceedings of the Conference on Computer-Supported Cooperative Work, 24-26 October 1994, Chapel Hill, North Carolina, ACM Press, New York, 1994c, pp. 101-112. Schmidt, Kjeld: ‘Mechanisms of interaction reconsidered’, in K. Schmidt (ed.): Social Mechanisms of Interaction, Computing Department, Lancaster University, Lancaster, UK, September 1994d, pp. 15-122. COMIC Deliverable D3.2. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Schmidt, Kjeld (ed.): Social Mechanisms of Interaction, Computing Department, Lancaster University, Lancaster, UK, September 1994e. COMIC Deliverable D3.2. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Schmidt, Kjeld, et al.: A ‘contrat sociale’ for CSCW systems: Supporting interoperability of computational coordination mechanisms, Centre for Cognitive Science, Roskilde University, Roskilde, Denmark, 1995. Working Papers in Cognitive Science and HCI. – WPCS-95-7. 472 Cooperative Work and Coordinative Practices Schmidt, Kjeld; and Tom A. Rodden: ‘Putting it all together: Requirements for a CSCW platform’, in D. Z. Shapiro; M. Tauber; and R. Traunmüller (eds.): The Design of Computer Supported Cooperative Work and Groupware Systems, North-Holland Elsevier, Amsterdam, 1996, pp. 157-176. – Originally published at the Workshop on CSCW Design, Schärding, Austria, 1-3 June 1993. Schmidt, Kjeld; and Carla Simone: ‘Coordination mechanisms: Towards a conceptual foundation of CSCW systems design’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 5, no. 2-3, 1996, pp. 155-200. Schmidt, Kjeld: ‘Of maps and scripts: The status of formal constructs in cooperative work’, in S. C. Hayne and W. Prinz (eds.): GROUP’97: Proceedings of the ACM SIGGROUP Conference on Supporting Group Work, 16-19 November 1997, Phoenix, Arizona, ACM Press, New York, 1997, pp. 138-147. Schmidt, Kjeld: ‘The critical role of workplace studies in CSCW’, in P. Luff; J. Hindmarsh; and C. C. Heath (eds.): Workplace Studies: Recovering Work Practice and Informing System Design, Cambridge University Press, Cambridge, 2000, pp. 141-149. Schmidt, Kjeld; and Carla Simone: ‘Mind the gap! Towards a unified view of CSCW’, in R. Dieng, et al. (eds.): Designing Cooperative Systems: The Use of Theories and Models. Proceedings of the 4th International Conference on the Design of Cooperative Systems (COOP 2000) [23-26 May 2000, Sophia Antipolis, France], IOS Press, Amsterdam etc., 2000, pp. 205221. Schmidt, Kjeld; and Ina Wagner: ‘Ordering systems in architectural design and planning: A discussion of classification systems and practices’, in G. C. Bowker; L. Gasser; and B. Turner (eds.): Workshop on Infrastructures for Distributed Collective Practice, 6-9 February 2002, San Diego, 2002a. Schmidt, Kjeld; and Ina Wagner: ‘Coordinative artifacts in architectural practice’, in M. BlayFornarino, et al. (eds.): Designing Cooperative Systems: A Challenge of the Mobility Age. [Proceedings of the 5th International Conference on the Design of Cooperative Systems (COOP 2002), 4-7 June 2002, Saint Raphaël, France], IOS Press, Amsterdam etc., 2002b, pp. 257-274. Schmidt, Kjeld: ‘Remarks on the complexity of cooperative work’, Revue des sciences et technologies de l’information. Série Revue d’intelligence artificielle (RSTI-RAI), vol. 16, no. 45. Paris, 2002a, pp. 443-483. Schmidt, Kjeld: ‘The problem with “awareness”: Introductory remarks on “Awareness in CSCW”’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 11, no. 3-4, 2002b, pp. 285-298. Schmidt, Kjeld; and Ina Wagner: ‘Ordering systems: Coordinative practices and artifacts in architectural design and planning’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 13, no. 5-6, 2004, pp. 349-408. Schmidt, Kjeld; Ina Wagner; and Marianne Tolar: ‘Permutations of cooperative work practices: A study of two oncology clinics’, in T. Gross, et al. (eds.): GROUP 2007: International Conference on Supporting Group Work, 4-7 November 2007, Sanibel Island, Florida, USA, ACM Press, New York, 2007, pp. 1-10. Schmidt, Kjeld: ‘Divided by a common acronym: On the fragmentation of CSCW’, in I. Wagner, et al. (eds.): ECSCW 2009: Proceedings of the 11th European Conference on ComputerSupported Cooperative Work, 7-11 September 2009, Vienna, Austria, Springer, London, 2009, pp. 223-242. Schmidt, Kjeld: ‘“Keep up the good work!”: The concept of “work” in CSCW’, in M. Lewkowicz, et al. (eds.): COOP 2010: 9th International Conference on the Design of Cooperative Systems, 18-21 May 2010, Aix-en-Provence, France, Springer, London, 2010. Schneider, Karin; and Ina Wagner: ‘Constructing the “dossier representatif”: Computer-based information-sharing in French hospitals’, Computer Supported Cooperative Work (CSCW): An International Journal, vol. 1, no. 4, 1993, pp. 229-253. Schonberger, Richard J.: Japanese Manufacturing Techniques: Nine Hidden Lessons in Simplicity, Free Press, New York, 1982. Schümmer, Till; and Stephan Lukosch: Patterns for Computer-Mediated Interaction, John Wiley & Sons, Chichester, England, 2007. References 473 Schütz, Alfred [A. Schutz]: ‘The problem of rationality in the social world’ (Economica, 1943). Text ed. by A. Brodersen. In A. Schutz: Collected Papers. Vol. II. Studies in Social Theory. Martinus Nijhoff, The Hague, 1964, pp. 64-90. Schütz, Alfred [A. Schutz]: ‘On multiple realities’ (Philosophy and Phenomenological Research, June 1945). Text ed. by M. Natanson. In A. Schutz: Collected Papers. Vol. I. The Problem of Social Reality. Martinus Nijhoff, The Hague, 1962, pp. 207-259. Schütz, Alfred [A. Schutz]: ‘The well-informed citizen: An essay on the social distribution of knowledge’ (Social Research, 1946). Text ed. by A. Brodersen. In A. Schutz: Collected Papers. Vol. II. Studies in Social Theory. Martinus Nijhoff, The Hague, 1964, pp. 120-134. Schütz, Alfred [A. Schutz]: Reflections on the Problem of Relevance (Manuscript, 1947-51). Text ed. by R. M. Zaner. Yale University Press, New Haven, 1970. Schütz, Alfred [A. Schutz]: ‘Common-sense and scientific interpretations of human action’ (Philosophy and Phenomenological Research, September 1953). Text ed. by M. Natanson. In A. Schutz: Collected Papers. Vol. I. The Problem of Social Reality. Martinus Nijhoff, The Hague, 1962, pp. 3-47. Schütz, Alfred [A. Schutz]: The Phenomenology of the Social World, Northwestern University Press, Evanston, Ill., 1967. Transl. from Der sinnhafte Aufbau der sozialen Welt; eine Einleitung in die Verstehende Soziologie, J. Springer, Wien, 1932. Transl. by G. Walsh and F. Lehnert. Searle, John R.: ‘Minds and brains without programs’, in C. Blakemore and S. Greenfield (eds.): Mindwaves: Thoughts on Intelligence, Identity and Consciousness, Basil Blackwell, Oxford, 1987, pp. 209-233. (Paperback ed., 1989). Searle, John R.: Minds, Brains and Science: The 1984 Reith Lectures, Pelican Books, London, 1989. Sellen, Abigail; and Richard H. R. Harper: The Myth of the Paperless Office, MIT Press, Cambridge, Mass., 2001. Selznick, Philip: ‘Foundations of the theory of organization’, American Sociological Review, vol. 13, 1948, pp. 25-35. Shanker, Stuart G. (ed.): Ludwig Wittgenstein: Critical Assessments: Volume Three: From the Tractatus to Remarks on the Foundations of Mathematics: Wittgenstein on the Philosophy of Mathematics, Croom Helm, London, 1986a. Shanker, Stuart G.: ‘Introduction: The portals of discovery’, in S. G. Shanker (ed.): Ludwig Wittgenstein: Critical Assessments: Volume Three: From the Tractatus to Remarks on the Foundations of Mathematics: Wittgenstein on the Philosophy of Mathematics, Croom Helm, London, 1986b, pp. 1-25. Shanker, Stuart G.: ‘Introduction: The nature of philosophy’, in S. G. Shanker (ed.): Ludwig Wittgenstein: Critical Assessments: Volume Four: From Theology to Sociology: Wittgenstein’s Impact on Contemporary Thought, Croom Helm, London, 1986c, pp. 1-28. Shanker, Stuart G.: Wittgenstein and the Turning-Point in the Philosophy of Mathematics, Croom Helm, London and Sydney, 1987a. Shanker, Stuart G.: ‘AI at the crossroads’, in B. P. Bloomfield (ed.): The Question of Artificial Intelligence: Philosophical and Sociological Perspectives, Croom Helm, London, etc., 1987b, pp. 1-58. Shanker, Stuart G.: ‘The decline and fall of the mechanist metaphor’, in R. Born (ed.): Artificial Intelligence: The Case Against, Routledge, London and New York, 1987c, pp. 72-131. Shanker, Stuart G.: ‘Wittgenstein versus Turing on the nature of Church's thesis’, Notre Dame Journal of Formal Logic, vol. 28, no. 4, October 1987d, pp. 615-649. Shanker, Stuart G.: ‘Turing and the origins of AI’, Philosophia mathematica, vol. 3–3, 1995, pp. 52-86. Shanker, Stuart G.: Wittgenstein’s Remarks on the Foundation of AI, Routledge, London, 1998. Shannon, Claude E.: ‘A mathematical theory of communication’, The Bell System Technical Journal, vol. 27, July, October 1948, pp. 379-423, 623-656. Shannon, Claude E.; and Warren Weaver: The Mathematical Theory of Communication, University of Illinois Press, Urbana, 1949. 474 Cooperative Work and Coordinative Practices Shapiro, Dan Z., et al.: ‘Visual re-representation of database information: the flight data strip in air traffic control’, in M. Tauber; D. E. Mahling; and F. Arefi (eds.): Cognitive Aspects of Visual Languages and Visual Interfaces, Elsevier Science, Amsterdam, 1994, pp. 349-376. Sharrock, Wes W.; and Robert J. Anderson: The Ethnomethodologists, Ellis Horwood Publishers, Chichester, 1986. Sharrock, Wes W.; and Rupert Read: Kuhn: Philosopher of Scientific Revolution, Polity Press, Cambridge, 2002. Sharrock, Wes W.; and Graham Button: ‘Plans and Situated Action ten years on’, The Journal of the Learning Sciences, vol. 12, no. 2, 2003, pp. 259-264. Sheil, Beau A.: ‘Coping with complexity’, Office: Technology and People, vol. 1, no. 4, January 1983, pp. 295-320. Shen, Weiming, et al. (eds.): CSCWD 2007: Computer Supported Cooperative Work in Design, IV: 11th International Conference, 26-28 April 2007, Melbourne, Australia, Springer, 2008. Shepherd, Allan; Niels Mayer; and Allan Kuchinsky: ‘Strudel: An extensible electronic conversation toolkit’, in T. Bikson and F. Halasz (eds.): CSCW’90: Proceedings of the Conference on Computer-Supported Cooperative Work, 7-10 October 1990, Los Angeles, Calif., ACM Press, New York, 1990, pp. 93-104. Shneiderman, Ben: ‘Direct manipulation: A step beyond programming languages’, IEEE Computer, vol. 16, no. 8, August 1983, pp. 57-69. Simon, Herbert A.: ‘The architecture of complexity’ (Proceedings of the American Philosophical Society, December 1962). In H. A. Simon: The Sciences of the Artificial. MIT Press, Cambridge, Mass., 1969, pp. 192-229. Simon, Herbert A.: ‘How big is a chunk?’ (Science, 1974). In H. A. Simon: Models of Thought. Yale University Press, New Haven and London, 1979, pp. 50-61. Simon, Herbert A.: Models of Thought, Yale University Press, New Haven and London, 1979. Simone, Carla; and Kjeld Schmidt (eds.): Computational Mechanisms of Interaction for CSCW, Computing Department, Lancaster University, Lancaster, UK, October 1993. COMIC Deliverable D3.1. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Simone, Carla: ‘Integrating co-operative work supports from CHAOS perspective (abstract)’, SIGOIS Bulletin, vol. 13, no. 4, 1993, p. 25. Simone, Carla; and Kjeld Schmidt (eds.): A Notation for Computational Mechanisms of Interaction, Computing Department, Lancaster University, Lancaster, UK, September 1994. COMIC Deliverable D3.3. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Simone, Carla, et al.: ‘A multi-agent approach to the design of coordination mechanisms’, in V. Lesser (ed.): Proceedings of the First International Conference on Multi-Agent Systems, San Francisco, Calif., USA, June 12-14, 1995, AAAI Press, Menlo Park, Calif., 1995a. Simone, Carla; Monica Divitini; and Kjeld Schmidt: ‘A notation for malleable and interoperable coordination mechanisms for CSCW systems’, in N. Comstock, et al. (eds.): COOCS’95: Conference on Organizational Computing Systems, 13-16 August 1995, Milpitas, California, ACM Press, New York, 1995b, pp. 44-54. Simone, Carla; and Stefania Bandini: ‘Compositional features for promoting awareness within and across cooperative applications’, in S. C. Hayne and W. Prinz (eds.): GROUP’97: Proceedings of the ACM SIGGROUP Conference on Supporting Group Work, 16-19 November 1997, Phoenix, Arizona, ACM Press, New York, 1997, pp. 358-367. Simone, Carla; and Monica Divitini: ‘Ariadne: supporting coordination through a flexible use of the knowledge on work processes’, Journal of Universal Computer Science, vol. 3, no. 8, 1997, pp. 865-898. Simone, Carla; and Kjeld Schmidt: ‘Taking the distributed nature of cooperative work seriously’, in: Proceedings of the 6th Euromicro Workshop on Parallel and Distributed Processing, 21-23 January 1998, Madrid, IEEE Computer Society Press, Los Alamitos, Calif., 1998, pp. 295301. Simone, Carla; and Monica Divitini: ‘Integrating contexts to support coordination: The CHAOS project’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 8, no. 3, September 1999, pp. 239-283. References 475 Simone, Carla; and Stefania Bandini: ‘Integrating awareness in cooperative applications through the reaction-diffusion metaphor’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 11, no. 3-4, 2002. Sluizer, S.; and P. Cashman: ‘XCP: An experimental tool for supporting office procedures’, in: Proceedings of the First International Conference on Office Automation, 17-19 December 1984, New Orleans, Louisiana, IEEE Computer Society Press, Silver Spring, Maryland, 1984, pp. 73-80. Smith, Adam: An Inquiry into the Nature and Causes of the Wealth of Nations, vol. 1-2, Printed for W. Strahan; and T. Cadell, in the Strand, London, 1776a. Smith, Adam: An Inquiry into the Nature and Causes of the Wealth of Nations, Printed for W. Strahan; and T. Cadell, in the Strand, London, 1776b. Text ed. by E. Cannan. – The Modern Library, New York, 1994. Smith, David Canfield, et al.: ‘Designing the Star User Interface’, Byte, vol. April, 1982a, pp. 242282. Smith, David Canfield, et al.: ‘The Star user interface: An overview’, in R. K. Brown and H. L. Morgan (eds.): AFIPS’82: Proceedings of the National Computer Conference, 7-10 June 1982, Houston, Texas, AFIPS Press, Arlington, Virginia, 1982b, pp. 515-528. Smith, H. T.; P. A. Hennessy; and G. A. Lunt: ‘An object-oriented framework for modelling organizational communication’, in J. M. Bowers and S. D. Benford (eds.): Studies in Computer Supported Cooperative Work. Theory, Practice and Design, North-Holland, Amsterdam, 1991, pp. 145-157. Smith, Hugh T.: ‘The requirements for group communication services’, in R. Speth (ed.): EUTECO ’88: European Teleinformatics Conference on Research into Networks and Distributed Applications, 20-22 April 1988, Vienna, Austria. Organized by the European Action in Teleinformatics COST 11ter, North-Holland, Brussels and Luxemburg, 1988, pp. 8995. Snedecor, George W.: ‘Uses of punched card equipment in mathematics’, The American Mathematical Monthly, vol. 35, no. 4, April 1928, pp. 161-169. Sombart, Werner: Der moderne Kapitalismus: Historisch-systematische Darstellung des gesamteuropäischen Wirtschaftslebens von seinen Anfängen bis zur Gegenwart: Band II: Das europäische Wirtschaftsleben im Zeitalter des Frühkapitalismus, Duncker & Humblot, München and Leipzig, 1916. (2nd ed.; 1st ed., 1902). – Deutscher Taschenbuch Verlag, München, 1987. Sørensen, Carsten: ‘The augmented bill of materials’, in K. Schmidt (ed.): Social Mechanisms of Interaction, Computing Department, Lancaster University, Lancaster, UK, September 1994a, pp. 221-236. COMIC Deliverable D3.2. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Sørensen, Carsten: ‘The CEDAC board’, in K. Schmidt (ed.): Social Mechanisms of Interaction, Computing Department, Lancaster University, Lancaster, UK, September 1994b, pp. 237-245. COMIC Deliverable D3.2. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Sørensen, Carsten: ‘The product classification scheme’, in K. Schmidt (ed.): Social Mechanisms of Interaction (ftp.comp.lancs.ac.uk), Computing Department, Lancaster University, Lancaster, UK, September 1994c, pp. 247-255. <ftp://ftp.comp.lancs.ac.uk/pub/comic> Sørgaard, Pål: ‘A cooperative work perspective on use and development of computer artifacts’, in: 10th Information Systems Research Seminar in Scandinavia (IRIS), 10-12 August 1987, Vaskivesi, Finland, 1987. Spear, Steven; and H. Kent Bowen: ‘Decoding the DNA of the Toyota Production System’, Harvard Business Review, September-October 1999, pp. 96-106. Speth, Rolf (ed.): EUTECO ’88: European Teleinformatics Conference on Research into Networks and Distributed Applications, 20-22 April 1988, Vienna, Austria. Organized by the European Action in Teleinformatics COST 11ter, North-Holland, Amsterdam etc., 1988. Sproull, Lee; and Sara Kiesler: Connections: New Ways of Working the the Networked Organization, The MIT Press, Cambridge, Mass., 1991. Standage, Tom: The Victorian Internet: The Remarkable Story of the Telegraph and the Nineteenth Century’s On-line Pioneers, Walker Publishing Co., New York, 1998. (Paperback ed., 2007). 476 Cooperative Work and Coordinative Practices Star, Susan Leigh; and James R. Griesemer: ‘Institutional ecology, “translations” and boundary objects: Amateurs and professionals in Berkeley’s Museum of Vertebrate Zoology, 1907-39’, Social Studies of Science, vol. 19, 1989, pp. 387-420. Star, Susan Leigh: ‘The structure of ill-structured solutions: Boundary objects and heterogeneous distributed problem solving’, in L. Gasser and M. Huhns (eds.): Distributed Artificial Intelligence, vol. 2, Pitman, London, 1989, pp. 37-54. Stefik, M., et al.: ‘WYSIWIS revised: Early experiences with multiuser interfaces’, ACM Transactions on Office Information Systems, vol. 5, no. 2, April 1987, pp. 147-167. Stinchcombe, Arthur L.: Creating Efficient Industrial Administrations, Academic Press, New York and London, 1974. Stinchcombe, Arthur L.: Information and Organizations, University of California Press, Berkeley, Calif., 1990. Storrs, Graham: ‘Group working in the DHSS large demonstrator project’, in P. Wilson; J. M. Bowers; and S. D. Benford (eds.): ECSCW’89: Proceedings of the First European Conference on Computer Supported Cooperative Work, 13-15 September 1989, Gatwick, London, London, 1989, pp. 102-119. Strauss, Anselm L.: ‘Work and the division of labor’, The Sociological Quarterly, vol. 26, no. 1, 1985, pp. 1-19. Strauss, Anselm L., et al.: Social Organization of Medical Work, University of Chicago Press, Chicago and London, 1985. Strauss, Anselm L.: ‘The articulation of project work: An organizational process’, The Sociological Quarterly, vol. 29, no. 2, 1988, pp. 163-178. Strauss, Anselm L.: Continual Permutations of Action, Aldine de Gruyter, New York, 1994. Styles, Keith: Working Drawings Handbook, Architectural Press, Amsterdam etc., 1995. (3rd edtion; 1st ed. 1982). Suchman, Lucy A.: ‘Systematics of office work: Office studies for knowledge-based systems. Digest’, in: Office Automation Conference, 5-7 April 1982, San Francisco, 1982, pp. 409-412. Suchman, Lucy A.: ‘Office procedure as practical action: Models of work and system design’, ACM Transactions on Office Information Systems, vol. 1, no. 4, October 1983, pp. 320-328. Suchman, Lucy A.; and Eleanor Wynn: ‘Procedures and problems in the office’, Office: Technology and People, vol. 2, 1984, pp. 133-154. Suchman, Lucy A.: Plans and Situated Actions: The Problem of Human-Machine Communication, Cambridge University Press, Cambridge, 1987. Suchman, Lucy A.: Notes on Computer Support for Cooperative Work, Dept. of Computer Science, University of Jyväskylä, Jyväskylä, Finland, May 1989. – WP-12. Suchman, Lucy A.; and Randall H. Trigg: ‘Understanding practice: Video as a medium for reflection and design’, in J. Greenbaum and M. Kyng (eds.): Design at Work: Cooperative Design of Computer Systems, Lawrence Erlbaum, Hillsdale, New Jersey, 1991, pp. 65-89. Suchman, Lucy A.: ‘Response to Vera and Simon's Situated action: A symbolic interpretation’, Cognitive Science, vol. 17, no. 1, January-March 1993a, pp. 71-75. Suchman, Lucy A.: ‘Technologies of accountability: On lizards and airplanes’, in G. Button (ed.): Technology in Working Order. Studies of work, Interaction, and Technology, Routledge, London and New York, 1993b, pp. 113-126. Suchman, Lucy A.: ‘Writing and reading: A response to comments on Plans and Situated Actions’, The Journal of the Learning Sciences, vol. 12, no. 2, 2003, pp. 299-306. Suchman, Lucy A.: Human-Machine Reconfigurations: Plans and Situated Actions, 2nd Edition, Cambridge University Press, New York, 2007. Sutherland, Ivan Edward: ‘Sketchpad: A man-machine graphical communication system’, in E. C. Johnson (ed.): SJCC’63: Proceedings of the Spring Joint Computer Conference, 21-23 May 1963, Detroit, Michigan, vol. 23, AFIPS Press, Santa Monica, California, 1963, pp. 329-346. Swade, Doron: The Cogwheel Brain: Charles Babbage and the Quest to Build the First Computer, Little, Brown & Co., London, 2000. – Abacus, London, 2001 (Paperback ed.). Swade, Doron: ‘The “unerring certainty of mechanical agency”: Machines and table making in the nineteenth century’, in M. Campbell-Kelly, et al. (eds.): The History of Mathematical Tables: From Sumer to Spreadsheets, Oxford University Press, Oxford, 2003, pp. 145-176. (Reprinted 2007). References 477 Swenson, K. D., et al.: ‘A business process environment supporting collaborative planning’, Collaborative Computing, vol. 1, no. 1, March 1994, pp. 15-24. Symon, Gillian; Karen Long; and Judi Ellis: ‘The coordination of work activities: Cooperation and conflict in a hospital context’, Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol. 5, no. 1, 1996, pp. 1-31. Syri, Anja: ‘Tailoring cooperation support through mediators’, in J. A. Hughes, et al. (eds.): ECSCW’97: Proceedings of the Fifth European Conference on Computer-Supported Cooperative Work, 7–11 September 1997, Lancaster, U.K., Kluwer Academic Publishers, Dordrecht, 1997, pp. 157-172. Tait, William W.: The Provenance of Pure Reason: Essays in the Philosophy of Mathematics and Its History, Oxford University Press, Oxford, 2005. Tang, John C.: ‘Findings from observational studies of collaborative work’, International Journal of Man-Machine Studies, vol. 34, 1991, pp. 143-160. Taylor, Frederick Winslow: The Principles of Scientific Management, Harper & Brothers, New York, 1911. – W. W. Norton & Co., New York, 1967. Taylor, Talbot J.: Mutual Misunderstanding: Scepticism and the Theorizing of Language and Interpretation, Duke University Press, Durham and London, 1992. Taylor, Talbot J.: Theorizing Language: Analysis, Normativity, Rhetoric, History, Pergamon, Amsterdam etc., 1997. Taylor, Talbot J.: ‘Language constructing language: the implications of reflexivity for linguistic theory’, Language Sciences, vol. 22, 2000, pp. 483-499. Thacker, Charles P., et al.: Alto: A personal computer, Xerox PARC, 7 August 1979. – CSL-7911. (Reprinted February 1984). <http://www.scribd.com/doc/2183225/Alto-A-Personal-Computer> Thacker, Charles P.: ‘Personal distributed computing: The Alto and Ethernet hardware’, in A. Goldberg (ed.): A History of Personal Workstations, Addison-Wesley, Reading, Mass., 1988, pp. 267-289. The OSI e-Infrastructure Working Group: Developing the UK’s E-infrastructure for Science and Innovation, Office of Science and Innovation (OSI), 2004. <http://www.nesc.ac.uk/documents/OSI/report.pdf> Thompson, James D.: Organizations in Action: Social Science Base of Administrative Theory, McGraw-Hill, New York, 1967. Thorngate, Warren: ‘Got a minute? How technology affects the economy of attention [Abstract of closing plenary talk]’, in W. A. Kellogg and S. Whittaker (eds.): CSCW 2000: ACM Conference on Computer Supported Cooperative Work, 2–6 December 2000, Philadelphia, Pennsylvania, USA, 2000. Tomlinson, Raymond S.: ‘The first network email’, 2001. <http://openmap.bbn.com/~tomlinso/ray/firstemailframe.html> Trevor, Jonathan; Tom A. Rodden; and Gordon Blair: ‘COLA: A lightweight platform for CSCW’, in G. De Michelis; C. Simone; and K. Schmidt (eds.): ECSCW’93: Proceedings of the Third European Conference on Computer-Supported Cooperative Work, 13-17 September 1993, Milano, Italy, Kluwer Academic Publishers, Dordrecht, 1993, pp. 15-30. Trevor, Jonathan; Tom A. Rodden; and Gordon Blair: ‘COLA: A lightweight platform for CSCW’, Computer Supported Cooperative Work (CSCW): An International Journal, vol. 3, no. 2, 1995, pp. 197-224. Turing, Alan M.: ‘On computable numbers with an application to the Entscheidungsproblem’, Proceedings of the London Mathematical Society, vol. 42 (Series 2), no. 3-4, NovemberDecember 1936, pp. 230-265. Turing, Alan M.: ‘Proposed electronic calculator (1945)’ (Report, 1945). In B. J. Copeland (ed.) Alan Turing’s Automatic Computing Engine: The Master Codebreaker’s Struggle to Build the Modern Computer. Oxford University Press, Oxford, 2005, pp. 369-454. – Written late 1945; submitted to the Executive Committee of the NPL February 1946 as ‘Report by Dr. A. M. Turing on Proposals for the Development of an Automatic Computing Engine (ACE)’. 478 Cooperative Work and Coordinative Practices Turing, Alan M.; and James H. Wilkinson: ‘The Turing-Wilkinson lecture series (1946-47)’ (Lectures, London, 1946-47). In B. J. Copeland (ed.) Alan Turing’s Automatic Computing Engine: The Master Codebreaker’s Struggle to Build the Modern Computer. Oxford University Press, Oxford, 2005, pp. 459-527. Turing, Alan M.: ‘Lecture on Automatic Computing Engine’ (London Mathematical Society, 20 February 1947). In B. J. Copeland (ed.) The Essential Turing: Seminal Writings in Computing, Logic, Philosophy, Artificial Intelligence, and Artificial Life plus The Secrets of Enigma. Oxford University Press, Oxford, 2004, pp. 378-394. Turnbull, David: ‘The ad hoc collective work of building Gothic cathedrals with templates, string, and geometry’, Science, Technology, & Human Values, vol. 18, no. 3, Summer 1993, pp. 315340. Turoff, Murray: ‘Delphi and its potential impact on information systems’, in R. R. Wheeler (ed.): FJCC’71: Proceedings of the Fall Joint Computer Conference, 16-18 November 1971, Las Vegas, Nevada, vol. 39, AFIPS Press, Montvale, New Jersey, 1971, pp. 317-326. Turoff, Murray: ‘Delphi conferencing: Computer-based conferencing with anonymity’, Technological Forecasting and Social Change, vol. 3, 1972, pp. 159-204. Turoff, Murray: ‘Human communication via data networks’, SIGCAS Computers and Society, vol. 4, no. 1, May 1973, pp. 15-24. Turoff, Murray: ‘Computer-mediated communication requirements for group support’, Journal of Organizational Computing, vol. 1, 1991, pp. 85-113. Uhlig, Ronald P.: ‘Human factors in computer message systems’, Datamation, vol. 23, no. 5, May 1977, pp. 120-126. Ure, Andrew: The Philosophy of Manufactures: or, an Exposition of the Scientific, Moral, and Commercial Economy of the Factory System of Great Britain, Charles Knight, London, 1835. – Frank Cass & Co., London, 1967. Urwick, Lyndall F.; and Edward F. L. Brech: The Making of Scientific Management. Volume I: Thirteen Pioneers, Sir Isaac Pitman And Sons, London, 1945. (Rev. ed., 1957). Urwick, Lyndall F.; and Edward F. L. Brech: The Making of Scientific Management. Volume II: Management in British Industry, Management Publications Trust, London, 1946. Urwick, Lyndall F.; and Edward F. L. Brech: The Making of Scientific Management. Volume III: The Hawthorne Investigations, Management Publications Trust, London, 1948. Vallee, Jacques F.: Group Communication through Computers. Volume 1: Design and Use of the FORUM System, Institute for the Future, Menlo Park, Calif., July 1974. – IFF Report R-32. van der Aalst, Wil M.P.: ‘Exploring the CSCW spectrum using process mining’, Advanced Engineering Informatics, vol. 21, 2007, pp. 191–199. Van der Spiegel, Jan, et al.: ‘The ENIAC: History, Operation, and Reconstruction in VLSI’, in R. Rojas and U. Hashagen (eds.): The First Computers: History and Architectures, The MIT Press, Cambridge, Mass., and London, 2000, pp. 121-178. Van Vleck, Tom: ‘The history of electronic mail’, 1 February 2001. Latest update, 25 May 2008. <http://www.multicians.org/thvv/mail-history.html> Vicente, Kim J.; and Jens Rasmussen: ‘Ecological interface design: Theoretical foundations’, IEEE Transactions on Systems, Man, and Cybernetics, vol. SMC-22, no. 4, 1992, pp. 589-606. Victor, Frank; and Edgar Sommer: ‘Supporting the design of office procedures in the DOMINO system’, in P. Wilson; J. M. Bowers; and S. D. Benford (eds.): ECSCW’89: Proceedings of the 1st European Conference on Computer Supported Cooperative Work, 13-15 September 1989, Gatwick, London, London, 1989, pp. 148-159. Victor, Frank; and Edgar Sommer: ‘Supporting the design of office procedures in the DOMINO system’, in J. M. Bowers and S. D. Benford (eds.): Studies in Computer Supported Cooperative Work. Theory, Practice and Design, North-Holland, Amsterdam, 1991, pp. 119-130. Von Glinow, Mary Ann; and Susan Albers Mohrman (eds.): Managing Complexity in High Technology Organizations, Oxford University Press, New York and Oxford, 1990. von Neumann, John: The First Draft Report on the EDVAC, Moore School of Electrical Engineering, University of Pennsylvania, 30 June 1945. – Contract No. W–670–ORD–4926 Between the United States Army Ordnance Department and the University of Pennsylvania. Text ed. by M. D. Godfrey. <http://qss.stanford.edu/~godfrey/vonNeumann/vnedvac.pdf> von Neumann, John: ‘The mathematician’, Works of the Mind, vol. 1, no. 1, 1947, pp. 180-196. References 479 von Savigny, Eike: Die Philosophie der normalen Sprache: Eine kritische Einführung in die ‘ordinary language philosophy’, Suhrkamp, Frankfurt a. M., 1974. (2nd ed., 1st. ed. 1969). von Savigny, Eike: Wittgensteins ‘Philosophische Untersuchungen’: Ein Kommentar für Leser. Band 1: Abschnitte 1-315, Vittorio Klostermann, Frankfurt a. M., 1994. (2nd ed.; 1st ed. 1988). Vygotskij, Lev Semenovič [L. S. Vygotsky]: ‘The instrumental method in psychology’ (Manuscript, 1930). In L. S. Vygotskij: The Collected Works of L. S. Vygotsky. Volume 3: Problems of the Theory and History of Psychology. Plenum Press, New York and London, 1997, pp. 85-89. – Theses of a talk read in 1930 at the N. K. Krupskaya Academy of Communist Education. Wagner, Ina; and Rainer Lainer: ‘Open planning: inspirational objects, themes, placeholders, and persuasive artefacts’, in: Proceedings Colloque Architecture des systèmes urbains, 5 July 2001, Université de Technologie de Compiègne, 2001. Waismann, Friedrich: ‘The nature of mathematics: Wittgenstein’s standpoint’ (Königsberg, 1930). Transl. from ‘Über das Wesen der Mathematik: Der Standpunkt Wittgensteins’. Transl. by S. G. Shanker. In S. G. Shanker (ed.) Ludwig Wittgenstein: Critical Assessments: Volume Three: From the Tractatus to Remarks on the Foundations of Mathematics: Wittgenstein on the Philosophy of Mathematics. Croom Helm, London, 1986, pp. 60-67. Waismann, Friedrich: Introduction to Mathematical Thinking: The Formation of Concepts in Modern Mathematics, Hafner Publishing Co., London, 1951. Transl. from Einführung in das mathematische Denken: Die Begriffsbildung der modernen Mathematik, Gerold u. Co., Wien, 1936. Transl. by T. J. Benac. Wakati, Osamu A.; and Richard M. Linde: Study Guide for the Professional Practice of Architectural Working Drawings, John Wiley & Sons, New York, 1994. (2nd ed., Student ed.). Wakefield, Edward: A View of the Art of Colonization, with present reference to the British Empire; in letters between a stateman and a colonist, John W. Parker, London, 1849. Waldrop, N. Mitchell: Complexity: The emerging science at the edge of order and chaos, Simon & Schuster, 1992. – Penguin Books, 1994 (Paperback ed.). Wang, Hao: From Mathematics to Philosophy, Routledge & Kegan Paul, London, 1974. Weber, Max: ‘Die wirtschaftlichen Unternehmungen der Gemeinden’ (28 September 1909). Text ed. by W. Schluchter. In M. Weber: Studienausgabe der Max-Weber-Gesamtausgabe. Mohr Siebeck, Tübingen, 1999, vol. I/8 (Wirtschaft, Staat & Sozialpolitik: Schriften und Reden, 1900-1912), pp. 127-130. Weick, Karl E.; and Karlene H. Roberts: ‘Collective mind in organizations: Heedful interrelating on flight decks’, Administrative Science Quaterly, vol. 38, 1993, pp. 357-381. Wertsch, James V.: Vygotsky and the Social Formation of Mind, Harvard University Press, Cambridge, Mass., and London, 1985. Wertsch, James V.: Voices of the Mind: A Sociocultural Approach to Mediated Action, Harvard University Press, Cambridge, Mass., 1991. White, Alan Richard: Attention, Basil Blackwell, Oxford, 1964. White, Alan Richard: The Philosophy of Mind, Random House, New York, 1967. White, Alan Richard: The Nature of Knowledge, Rowman and Littlefield, Totowa, New Jersey, 1982. Whitehead, Alfred North; and Bertrand Russell: Principia Mathematica, Cambridge University Press, Cambridge, 1910. (2nd ed., 1927; paperback ed. ‘to *56’, 1962). Whitley, Richard: ‘Cognitive and social institutionalization of scientific specialties and research areas’, in R. Whitley (ed.): Social Processes of Scientific Development, Routledge & Kegan Paul, London, 1974, pp. 69-95. Wiedemann, Thomas: Greek and Roman Slavery, Croom Helm, London, 1981. – Routledge, New York, 1988. Wieser, C. Robert: Cape Cod System and demonstration, MIT Lincoln Laboratory, Division 6, Cambridge, Mass., 13 March 1953. – Project Whirlwind Limited Memorandum VI - L-86. <http://dome.mit.edu/handle/1721.3/41510> Wieser, C. Robert: ‘The Cape Cod System’, IEEE Annals of the History of Computing, vol. 5, no. 4, April-June 1985, pp. 362-369. 480 Cooperative Work and Coordinative Practices Wilczek, Stephan; and Helmut Krcmar: ‘Betriebliche Groupwareplattformen’, in G. Schwabe; N. Streitz; and R. Unland (eds.): CSCW-Kompendium: Lehr- und Handbuch zum computerunterstützten kooperativen Arbeiten, Springer, Heidelberg, 2001, pp. 310-320. Wild, Ray: Mass-production Management: The Design and Operation of Production Flow-line Systems, John Wiley & Sons, London, 1972. Williams, Meredith: Wittgenstein, Mind and Meaning: Toward a Social Conception of Mind, Routledge, London and New York, 1999. Williams, Michael R.: ‘Early calculation’, in W. Aspray (ed.): Computing Before Computers, Iowa State University Press, Ames, Iowa, 1990, pp. 3-58. Williams, Michael R.: ‘Difference engines: From Müller to Comrie’, in M. Campbell-Kelly, et al. (eds.): The History of Mathematical Tables: From Sumer to Spreadsheets, Oxford University Press, Oxford, 2003, pp. 123-144. (Reprinted 2007). Williamson, Oliver E.: Markets and Hierarchies: A Transactional and Antitrust Analysis of the Firm, Free Press, New York, 1975. Williamson, Oliver E.: ‘Transaction-cost economics: The governance of contractual relations’, Journal of Law and Economics, vol. 22, no. 2, October 1979, pp. 233-261. Williamson, Oliver E.: ‘The economics of organization: The transaction cost approach’, American Journal of Sociology, vol. 87, no. 3, November 1981, pp. 548-577. Wilson, Paul: Computer-Supported Cooperative Work. An Introduction, Intellect, Oxford, England, 1991. Winch, Peter: The Idea of a Social Science and its Relation to Philosophy, Routledge & Kegan Paul, London, 1958. (Second ed., 1990). Winograd, Terry: ‘A language/action perspective on the design of cooperative work’, in H. Krasner and I. Greif (eds.): CSCW’86: Proceedings. Conference on Computer-Supported Cooperative Work, Austin, Texas, 3-5 December 1986, ACM Press, New York, 1986, pp. 203220. Winograd, Terry; and Fernando Flores: Understanding Computers and Cognition: A New Foundation for Design, Ablex Publishing Corp., Norwood, New Jersey, 1986. Wittgenstein, Ludwig: Wittgenstein and the Vienna Circle: Conversations Recorded by Friedrich Waismann (1929-32). Transl. by J. Schulte and B. McGuiness. Text ed. by B. McGuiness. Basil Blackwell, Oxford, 1979. Wittgenstein, Ludwig: Philosophical Remarks (Manuscript, 1930). Transl. from Philosophische Bemerkungen. Transl. by R. Hargreaves and R. White. Text ed. by R. Rhees. Basil Blackwell, Oxford, 1975. Wittgenstein, Ludwig: Philosophical Grammar (Manuscript, 1931-34). Transl. by A. Kenny. Text ed. by R. Rhees. Basil Blackwell, Oxford, 1974; Paperback ed. 1980. Wittgenstein, Ludwig: Philosophy for Mathematicians: Wittgenstein’s Lectures, 1932-33. From the Notes of Alice Ambrose (1932-33). Text ed. by A. Ambrose. In L. Wittgenstein: Wittgenstein’s Lectures: Cambridge, 1932-1935. From the Notes of Alice Ambrose and Margaret Macdonald. Basil Blackwell, Oxford, 1979, pp. 203-225. Wittgenstein, Ludwig: The Yellow Book: Wittgenstein’s Lectures and Informal Discussions during dictation of the Blue Book 1933-34. From the Notes of Alice Ambrose (1934-35). Text ed. by A. Ambrose. In L. Wittgenstein: Wittgenstein’s Lectures: Cambridge, 1932-1935. From the Notes of Alice Ambrose and Margaret Macdonald. Basil Blackwell, Oxford, 1979, pp. 41-73. Wittgenstein, Ludwig: Remarks on the Foundations of Mathematics (Manuscript, 1937-44). Transl. from Bemerkungen über die Grundlagen der der Mathematik. Transl. by G. E. M. Anscombe. Text ed. by G. H. von Wright; R. Rhees; and G. E. M. Anscombe. Basil Blackwell Publishers, Oxford, 1978. Wittgenstein, Ludwig: Lectures on the Foundations of Mathematics, Cambridge, 1939: From the Notes of R. G. Bosanquet, Norman Malcolm, Rush Rhees, and Yorick Smythies (1939). Text ed. by C. Diamond. The University of Chicago Press, Chicago and London, 1976. Wittgenstein, Ludwig: Philosophical Investigations (Manuscript, 1945-46). Transl. from Philosophische Untersuchungen. Transl. by G. E. M. Anscombe; P. M. S. Hacker; and J. Schulte. Text ed. by P. M. S. Hacker and J. Schulte. In L. Wittgenstein: Philosophical Investigations. Blackwell Publishers, Oxford, 4rd ed., 2009, pp. 1-181. – TS 227: Previously published as Philosphical Investigations, Part 1. References 481 Wittgenstein, Ludwig: Zettel (Manuscript, 1945-48). Transl. by G. E. M. Anscombe. Text ed. by G. E. M. Anscombe and G. H. von Wright. Basil Blackwell Publishers, Oxford, 1967, 2nd edition 1981. Wittgenstein, Ludwig: Philosophical Investigations (Manuscript, 1945-49). Transl. from Philosophische Untersuchungen. Transl. by G. E. M. Anscombe. Basil Blackwell Publishers, Oxford, 1958. Wittgenstein, Ludwig: Remarks on the Philosophy of Psychology. Volume I (Manuscript, 194649). Transl. by C. G. Luckhardt and M. A. E. Aue. Text ed. by G. H. von Wright and H. Nyman. Basil Blackwell Publisher, Oxford, 1980. Wittgenstein, Ludwig: On Certainty (Manuscript, 1949-51). Transl. by D. Paul and G. E. M. Anscombe. Text ed. by G. E. M. Anscombe and G. H. von Wright. Basil Blackwell Publishers, Oxford, 1967, 2nd ed. 1981. Womack, James P.; Daniel T. Jones; and Daniel Roos: The Machine that Changed the World: The Story of Lean Production, Rawson Associates, New York, 1990. – Harper Collins Publishers, New York, 1991. Woo, Carson C.; and Frederick H. Lochovsky: ‘Supporting distributed office problem solving in organizations’, ACM Transactions on Office Information Systems, vol. 4, no. 3, July 1986, pp. 185-204. Woods, David D.: ‘Commentary: Cognitive engineering in complex and dynamic worlds’, International Journal of Man-Machine Studies, vol. 27, no. 5-6, November-December 1987, pp. 571-585. Woods, David D.: ‘Coping with complexity: the psychology of human behavior in complex systems’, in L. P. Goodstein; H. B. Andersen; and S. E. Olsen (eds.): Tasks, Errors and Mental Models. A Festschrift to celebrate the 60th birthday of Professor Jens Rasmussen, Taylor & Francis, London, 1988, pp. 128-148. Wynn, Eleanor H.: Office Conversation as an Information Medium, Ph.D. dissertation, University of California, Berkeley, 1979. Wynn, Eleanor H.: ‘Taking practice seriously’, in J. Greenbaum and M. Kyng (eds.): Design at Work: Cooperative Design of Computer Systems, Lawrence Erlbaum, Hillsdale, New Jersey, 1991, pp. 45-64. Yates, JoAnne: Control through Communication: The Rise of System in American Management, Johns Hopkins University Press, Baltimore and London, 1989. Zerubavel, Eviatar: Patterns of Time in Hospital Life: A Sociological Perspective, University of Chicago Press, Chicago and London, 1979. Zimmerman, Don Howard: Paper Work and People Work: A Study of a Public Assistance Agency, PhD diss., University of California, Los Angeles, 1966. Zimmerman, Don Howard: ‘Record-keeping and the intake process in a public welfare agency’, in S. Wheeler (ed.): On Record: Files and Dossiers in American Life, Russell Sage Foundation, New York, 1969a, pp. 319-354. Zimmerman, Don Howard: ‘Tasks and troubles: The practical bases of work activities in a public assistance agency’, in D. A. Hansen (ed.): Explorations in Sociology and Counseling, Houghton-Mifflin, New York, 1969b, pp. 237-266. Zisman, Michael D.: Representation, Specification and Automation of Office Procedures, Ph.D. dissertation, Dept. of Decision Sciences, The Wharton School, Univ. of Pennsylvania, PA, 1977. 483 Index