This set of documentation is intended to provide a critical overview of what is meant by the terms computer supported cooperative work (CSCW) and groupware. As such, it is intended to serve as an advanced introduction and analysis of these constructs. For basic information on how these terms have obtained their present connotations, the reader is referred to the core literature listed in the Literature Guide
The discussion of the label "CSCW" breaks down into two aspects, based on a dualistic usage of the term. First, "CSCW" denotes a phenomenon or subject area of interest. Second, "CSCW" denotes a research field comprised of practitioners from a variety of conventionally-recognized disciplines. The discussion of the label "groupware" has essentially only one aspect -- the denotation of those instrumentalities through which CSCW (the phenomenon) is realized and toward which CSCW (the field) addresses itself.
This overview will proceed as follows. First, the basic canonical definitions for these terms will be introduced to provide fundamental orientational guidance. Second, the phenomenon of "CSCW" will be analyzed with respect to previous attempts at resolving its connotations into something more like a denotation. Third, the instrumentalities typically subsumed under the label "groupware" will be similarly analyzed. Fourth, the field of "CSCW" will be addressed with respect to points arising in the foregoing discussions.
The bulk of this discussion will introduce and wrestle with the problems in trying to positively define and delineate computer supported cooperative work and groupware. To 'bootstrap' the discussion, this section will provide summary accounts of the canonical definitions.
The label computer-supported cooperative work (CSCW) was coined by Irene Greif and Paul Cashman in 1984 as a marketing tag for a vision of integrated office IT support -- "...A shorthand way of referring to a set of concerns about supporting multiple individuals working together with computer systems." (Bannon & Schmidt, 1989, p. 358). By 1989, The Institute for the Future's Robert Johansen (one of this field's most perceptive and persistent promoters) would list CSCW among a set of 14 terms in use to describe the emerging phenomenon. The other 13 were:
Not all these candidate terms were promoted as applicable to the entirety of the foreseen world of networked team knowledge work. In any case, CSCW was the one which came to serve as the general label for both the phenomenon of interest (group-oriented applications of information technology -- IT) and the research and development field dedicated to studying it. Proponents question the precise boundaries of the term (in both the senses mentioned), though none question the importance of the issues addressed therein.
Work under this rubric has generated a significant body of research, development, and trade literature. The reader wishing to more deeply explore the origins, themes, and developments in this area is recommended to Baecker (1992); Marca & Bock (1992); Greenberg (1991); Greif (1988); Olson (1989); and Johansen (1988) -- all excellent introductory overviews to this area and the issues it covers. For analytical overviews of how group IT applications have actually been applied in real organizations, one can review the substantive literature which began appearing with Bullen & Bennett (1990a; 1990b) and Groupware Resources Guide (Groupware Users' Project, 1990).
As we pass the midpoint of the 1990's, these remain the primary foundational references in the field, because attention to basic theoretical issues began to wane circa 1990-'91, when IT applications (e.g., Lotus Notes) started achieving market penetration on the basis of group / organizational foci. The proceedings from the annual CSCW conferences also provide timely and valuable sources of material at both introductory and advanced levels. These conferences alternate location between North America (even-numbered years) and Europe (odd-numbered years).
The themes and topics addressed in CSCW are the subject of work which proceeds under a variety of taxonomic labels. Related terminology for this general topical terrain includes: business process reengineering (BPR), computer-mediated communications (CMC), computer supported cooperative learning (CSCL), coordination science, organizational learning, organizational (re-)design, and participatory design (PD). Some of these areas represent specialized domain applications of CSCW interests (e.g., CSCL) and some represent related fields whose focus diverges from that of CSCW as analyzed later in this document (e.g., CMC).
The term groupware is conventionally used to denote those products or applications supporting work groups, but this term is not universally accepted. The word was coined in 1978 by Johnson-Lenz & Johnson-Lenz (cf. 1992, p. 130) to denote:
Intentional GROUP processes and procedures to achieve specific purposes
plus
softWARE tools designed to support and facilitate the group's work
(Hiltz & Turoff, 1992, p. 69)
As such, the label was originally intended to subsume both IT artifacts and workplace social systems. Coleman (1992, p. xiii) states groupware is "computer-mediated collaboration that increases the productivity or functionality of a person-to-person(s) interactive process." In a presentation to CSCW '88 (cf. Johnson-Lenz & Johnson-Lenz, 1992, p. 130), Doug Engelbart characterized groupware in terms of a co-evolving human-tool system. As time went on, Engelbart maintained this broad view, calling groupware "...a strategic means to an important end: creating high-performance human organizations" (1992, p. 77), but clearly associated groupware with the "tool-based" side of his capability infrastructure (1992, p. 78).
Over the years, the scope of the term has narrowed in conventional use to more specifically connote the technical artifacts or tools themselves. The typical current definition for groupware now runs something like this: "computer-based systems that support groups of people engaged in a common task (or goal) and that provide an interface to a shared environment." (Ellis, Gibbs & Rein: 1991, p. 40; 1993, p. 10). This shift toward narrow ascription of the term to the technical aspects of IT-supported collaboration has resulted in the common differentiation between CSCW and groupware. For example, Bannon and Schmidt (1989) discriminate between groupware and group work issues, as do Grudin (1991) and Johansen (1988). This narrower usage has not resulted in any appreciable denotational precision, and the label remains "...about as useful a term as 'singleware'." (Dyson, 1992, p. 10)
Throughout the remainder of this Webspace, the conventional distinction between (a) CSCW as a phenomenon / topical area and (b) groupware as the IT artifacts relating to CSCW will be maintained.
There are a variety of paths which have led to this area currently labeled CSCW, and one might select any of a number of places to begin tracing its history. As it concerns the confluence of IT with work groups, CSCW clearly addresses the intersection of technical and social interests. We shall see that to the extent the field has defined itself, it is neither wholly technical nor wholly social. The diverse historical threads now intertwined in CSCW cannot be extracted from their overall context and analyzed in isolation. Some trends (e.g., the proliferation of IT in working life) are so ubiquitous they fail to uniquely specify CSCW. Some pertinent allied fields (e.g., anthropology, social psychology) overlap with respect to subject matter, but have not directly promoted the rise of CSCW as a research area. For the purposes of this discussion, I shall outline only enough of these contributory trends to sketch CSCW's historical background.
Most broadly, CSCW can be seen as an result of three key background factors (Teknikstöd för Grupparbete, 1990) which are apparent from the conventional vantage of industrial business management:
These factors translate into combined demands for (1) coordination among workers and (2) management of increasingly important informational resources through the use of telecommunications and data processing. Early work on group IT support directly addressed these demands. The original 'coordination systems' were devised as a means for controlling complexity in software programming projects (e.g., MONSTR -- Cashman & Holt (1980)), and the office information systems of the late 1970's were geared toward common access to work documentation. In both cases, the orientational focus for IT interventions was the enhancement of tangible production activities (e.g., generation of software code or documents). In effect, the initial applications of IT for team support were the computerized equivalent to mechanization on the shop floor. The obvious motivation was what Habermas would call a technical interest, insofar as they addressed worker interactivity only "...insofar as it appears within the behavioral system of instrumental action." (Habermas, 1971, p. 195)
But the above is only one perspective of the historical threads which knit together as CSCW and groupware. It should be self-evident that an area whose very labels connote technical intervention (i.e., computer supported...; groupware) would also be historically contingent upon technological advances. The key developments in this case include (e.g.) the phenomenon of the Internet, the evolution of desktop computing, office information systems, and the evolution of feasible local area networks (LANs). More to the point were the advances in the utility of data networks (computer support) for interpersonal communication (the essential substrate for cooperative work). The landmark efforts of Starr Roxanne Hiltz and Murray Turoff laid the foundations for distributed IT usage back in the 1970's (cf. The Network Nation, 1993), and they have been appropriately canonized within the cognate area of computer-mediated communication (CMC).
I shall pause long enough to lay out two of the specific lines of work which have laid the foundations for current research and development in CSCW and groupware, operating from a Habermasian technical interest. One of them concerns information technology itself, and the other concerns an analytical orientation to the social / organizational nature of the workplace.
For better or for worse, the joint technological / social innovations associated with these themes is upon us. In the next section, I shall briefly outline the darker side of the ongoing shift.
Owing to maintenance of the Industrial Revolution's predominant technical knowledge interest, we find that some initial effects of group IT deployment resemble the outcomes of factory automation. If within this technical orientation one accepts an analogy between flows of materiel and of information, the above-cited demands for task coordination and information management parallel the demands for production control and resource management which arose during the Industrial Revolution. This is the extent of how far many conventional managers and analysts can conceptualize the horizon of imminent change. As a result, it should come as no surprise that there are already examples of detrimental outcomes analogous to those wrought by the excesses of scientific management (Taylor, 1947). Both task coordination and information resource management, when framed too strictly with regard to the precepts and presumptions of the 'Second Wave' (Toffler & Toffler, 1980; 1990) Industrial Revolution, have sometimes netted negative results from extreme exercise of their purported 'magic bullets' (respectively: workflow / BPR and data pooling / security).
The reason for such unfortunate outcomes largely derives from the fact that those mired in the conventional mindset just "don't get it." Information and knowledge are not so rigidly finite as materials; they can be shared to synergistic advantage in relative defiance of the spatiotemporal constraints applicable to physical assets. This changes the loci and foci of value-adding opportunities in information / knowledge work. For example, such work lends itself to collaborative interactivity at both the point source of value creation and subsequent value-enhancing stages. In addition, the number of "chefs" who can synchronously (or 'virtually' synchronously) "cook the soup" can vary over a greater range than (e.g.) the number of assembly workers who can simultaneously bolt parts onto a chassis. These (and other) novel affordances of knowledge work diminish reliance upon the Industrial Revolution's more simplisticly linear, 'downstream-flowing' process channels through which extant materials are manipulated into more profitable forms.
This greatly complicates the topography of work processes involved in producing a given output. In turn, this heightened complexity confounds many traditional techniques for activity coordination. As if that weren't problem enough, the emergent knowledge workstyle is transforming the very scope of such work processes and their products. Collaborative practices in the 'second wave' industrial paradigm (cf. Toffler & Toffler, 1980; 1990) evolved around progressive and incremental addition of conventionally-determinable exchange value to pre-given finite elements. In contrast, 'third wave' collective activity is increasingly moving onward / outward to create new types and sources of value and valuation. In other words, the very dimensions delimiting a work process topography are now subject to change, taking the above-mentioned confounding complexity to a new order of magnitude.
The current transformations are not limited to basic production work, either. IT is now being applied to mediate the functions previously relegated to management themselves -- allocation of resources, scheduling, and the like. The results are opportunities for increased workplace cooperation combined with decreased workplace stratification. Insofar as cooperation is as beneficial across as within enterprise boundaries and stratification (if preserved overlong) can have counterproductive effects, these opportunities are increasingly attractive in this period of broadening markets and flexible enterprises. Or at least that's the way management consultants like to characterize the times -- as if the techniques and technologies they promote were precision-tailored responses to events which are no more than trends or fads in the course of conventional 'business' or 'organization'.
Framing one's attitude to these transitions and transformations with respect to potentially obsolescent notions of 'business' and/or 'organization' is 'shortsighted' in two senses. First, it is overly focussed on the near term and thereby neglects ramifications which will become manifest when the present 'wave of change' is spent. By the mid-1990's, some organizations whose 'reengineering' efforts led to 'downsizing' of their key second-wave resources (wage labor and materiel) were confronted with resultant shortages in the key third-wave resources (resident 'knowledge' and 'knowledge skills') upon whose recognition the reengineering was promoted in the first place.
The second sense of 'shortsightedness' is a myopia which excludes serious consideration of the disturbing notion that it is the momentum of the techniques and techologies which will overcome the inertia of a second wave worldview, and not the other way around. The rise of the Internet is better explained by 'demand pull' in the user population than by 'supply push' on the part of its creators (the U.S. military/industrial complex). Innovations such as internal and external market networks, as well as less rigid and hierarchical workplace structures, are better explained as organic outcomes (as opposed to planned payoffs) of the collaborative affordances intrinsic to information technologies. To paraphrase Haldane's line on the queerness of the universe: the situation is not just outside the immediate control of the coordinators and managers -- it is outside the scope of their possible control.
By the time of this writing, CSCW had become a victim of its focal phenomenon's self-evident criticality. The proliferation of local and global data networks and the arrival of commercial group IT products combined to shift attention from delineating the scope and focus of enquiry to divining the marketplace. The early enthusiasm for theorization in CSCW circles began to wane about the time Lotus Notes appeared. This correlates with my personal observation that theory began losing out to technical details after the CSCW'90 (Los Angeles) and ECSCW'91 (Amsterdam) conferences. By November 1996 (at CSCW'96), the situation was such that during a plenary discussion on CSCW's future one panelist bemoaned the lack of any "indigenous theory" for the field. In the best case, this represents a belated recognition of the field's being highjacked by market forces some 5 years earlier. In the worst case, this represents an amnesia toward the significant work done before this highjacking occurred.
The remainder of this document shall therefore backtrack to the halcyon days of CSCW's explicit uncertainties about itself and its subject matter. We shall examine some of the themes whose evolution was subordinated to the pressures of market forces within the development community and to the exigencies of cyclical turnover in graduate students within academia.
In this section, we shall explore the ways in which the subject matter of the CSCW discipline has been delimited and dissected. It will be in a much later section that we shall try to delimit the boundaries of the discipline itself.
In trying to circumscribe "CSCW", some authors have given their attention to discussing work, while others have concentrated on IT products. This divergence of focus (i.e., products versus work processes) has been widely recognized and discussed. Bannon & Schmidt (1989, p. 359) distinguish between "CS" and "CW" (respectively) in illustrating this divide, while Grudin (1991) contrasts (respectively) "groupware" with "CSCW". A closely related distinction is geographically delineated between the product-oriented, technical tenor of much American work and the work-oriented, social emphasis more common in Europe (Grudin, 1991). At its extreme, this has resulted in a bipolar functionalist / philosophical divide jokingly (but accurately) dubbed: "Lotus Notes versus Notes on the Lotus." Whitaker, Östberg & Essler (1989) try to overcome the exclusionary tone of the more extreme analyses when they distinguish between CScw (Computer Support for cooperative work -- the technical perspective) and csCW (computer support for Cooperative Work -- the social / organizational perspective). This terminological 'gimmick' was invoked to highlight the dichotomy between relative emphases found in the literature, without obscuring the linkage between both sides which is self-evidently a key to this area's context and allure. This last nomenclature for the distinction will be employed henceforth in this document.
Largely due to these differential foci, CSCW is impossible (and groupware very difficult ...) to precisely circumscribe (cf. Howard,1986; Bannon & Schmidt,1989; Grudin,1991); and Robinson: 1989;1991). The central notions of computer support and cooperative work provide a basic agenda for discussion, but to date there has been little progress in bringing together a unified positive definition for their focal confluence. Bob Johansen (1991) has suggested such ongoing debate is not cause for giving up or turning cynical -- indeed, he claimed the combination of motivated interest and confusion which typifies the CSCW area is a sure sign of its importance. Although there is probably considerable wisdom in this position, it is no excuse for sitting still. In an attempt at some definitional progress, the following sections will introduce some contrasts to help delimit the "boundaries" circumscribing the subject matter of CSCW and groupware studies.
The term CSCW becomes no more lucid when a purely 'cooperative' character of the work process is taken as the focal point for analysis. Much discursive effort was expended in this direction during the mid- to late-1980's, especially among European writers. After an extensive review, Bannon & Schmidt (1989) conclude "...the term 'cooperative work' is the general and neutral designation of multiple persons working together to produce a product or service." (p. 362) Their lack of further specificity is understandable; "cooperation" becomes a very problematical subject under closer scrutiny. Schmidt (1988) is able to distinguish three interlinked "forms of cooperation" (pp. 10-28), but such detail requires strictly adhering to Marx's production-oriented view and explicitly minimizing allusion to general social interactions (cf. his discussion on p. 9) among cooperating parties.
Sørgaard (1987) attempted an analytical definition of what is meant by "cooperative work". This he bases on four criteria for defining a work situation as being cooperative in nature:
These criteria emphasize a balance of equality among cooperating actors and autonomy of their cooperative efforts with respect to any subsuming authority. Sørgaard (1988b) then circumscribed CSCW as applying IT "in a way which supports the cooperative aspects of the work" (pp. 2-3). This line of analysis has attracted more than its share of dismissive hostility. Howard (1987, p. 175-176) criticized this " 'Mom-and-apple-pie' connotation" of cooperation for being "...not merely a description of the way work is but a prescription for the way it ought to be." Such alleged idealism with respect to 'cooperation' has continued to attract cautionary comments in the literature (e.g., Kyng, 1991, p. 69). Even more striking is the fact that this outlook has been criticized as the result of a particular personal bias. Bødker et al. (1988, pp. 378-380) suggested such viewpoints simply reflect the privileged character of the research groups in which CSCW authors themselves participate (e.g., the teams at Xerox PARC), making this orientation too simplistic or limited to serve as a general exemplar. At this point, the responses to Sørgaard had become notably 'catty' in tone. Given that it would be construable as similarly 'catty' to note the irony in such accusations of 'idealism' coming from sources (the Aarhus PD community) themselves infamous for starry-eyed hand-waving (Can you say "workplace democracy"?), we shall pursue this line of discussion no further. Suffice it to say that attempts to positively define "cooperation" effectively ended with Sørgaard.
This does not mean that attempts to circumscribe "cooperation" wholly disappeared. Others, working from a broader vantage, addressed "cooperation" in terms of more general social or organizational factors. Howard (1987) claims "cooperative" is biased in tacitly ignoring conflicts inherent in work organizations. The common surrogate "collaborative" is no better because it insinuates unwarranted autonomy for individual participants (i.e., it implies voluntary choice on the part of persons empowered to choose). On a more figurative note, "collaboration" can also connote passive compliance or active cooperation with an enemy (cf. Bannon & Schmidt, 1989, p. 362). While this connotation rings true with many workers' tolerance of IT interventions, it has not been pursued. Howard goes on to select "collective" as a more appropriate term. He, Bannon, and Schmidt were in general accord on their usage of this term "collective" to connote a "fusion of the members ... into a whole" (Ibid.) and an increased measure of accountability for the unit thus created (Howard, 1987, pp. 177-178). By this point, the analytical efforts were widely taken to be waxing overly pedantic, and attempts to pin down a proper label for the focal work style effectively faded out.
It is naive to presumptively link collaborative work activity with a benign socio-political milieu. Grudin (1988), while suggesting "there may be a shift toward greater egalitarianism in the workplace..." (p. 92), never asserts this as a mandatory effect of CSCW system implementation. Indeed, much of his discussion on conditions for groupware acceptance concerns vertically stratified hierarchical organizations. More definitively, Bannon & Schmidt (1989) state "(t)he concept of cooperative work does not imply a particular degree of participation or self-determination on the part of the workers, nor a particularly democratic management style." (p. 362)
It is similarly naive to presume that the introduction of tools for collaboration will promote such a benign milieu. Whitaker, Östberg & Essler (1989) suggest that groupware products' intrinsic presumptions may occasionally violate cultural mores as well as political ideals. It would appear that this precaution is borne out in practice. Caracik & Grantham (1988) describe users' rejection of a coordination system on the grounds of its perceived negative impacts on autonomy and equality. Bullen & Bennett (1990a; 1990b) note the social impacts of groupware implementations have apparently affected their assessment and continuance, and their findings of conflicting customer attitudes (between workers and managers) support Grudin's (1988) cautionary emphasis on disparities of empowerment.
The issues become no clearer if one shifts from qualities of work spanning a bloc to the topographical scope of the bloc itself. CSCW is not delineated with strict regard to organizational boundaries. Skeptics have long viewed CSCW as a repetition of the office automation fad of the late 1970's, emphasizing integrated production support for entire organizational units (e.g., Wohl, 1989). With specific regard to strict organizational delineation, such comparisons are unfounded. The last decade's drive toward integration of LAN- and WAN-based IT has largely been a tale of either overcoming or succumbing to hardware, software, and other differentials among (and within) units at all levels of organizational constitution.
On the other hand, there is no necessity that consideration of CSCW be properly constrained to a single organization. Toffler's (1990) interconnected "power mosaics" (flexible collaborative networks, successors to monolithic enterprises) rely on inter-organizational communication and coordination, and they will provide a major impetus to groupware proliferation. Suomi (1989) and Hart and Estrin (1990) provide general overviews of IT systems for coordinating operations across organizational boundaries, while Engelbart (1990) describes an example of "knowledge domain interoperability" from the aerospace industry, spanning some 6,000 separate companies. In other words, whatever "collective" exemplifies CSCW need not correspond to any specific region of the official organizational map.
De Michelis (1990) differs from both the extreme CScw and the extreme csCW viewpoints in attempting to address cooperative work in such a manner as to provide a means for discussing specific software applications deriving therefrom. His approach concentrates on the manner of cooperative activity, rather than on a comprehensive definition of cooperation itself. In fact, De Michelis does not attempt to define what he means by "cooperation". Instead, he notes a trend toward the use of task-directed groups in modern enterprises and claims those groups are "...defined by the pattern of commitments that group members make with each other and with third parties." (p. 2) Having established this focus, De Michelis proceeds to delineate three different categories of cooperation:
Based on this trinary distinction, De Michelis then proceeds to discuss specific types of support systems developed to date. He does not address a general process of "cooperation", as was the goal with Sørgaard's analysis or the terminological debates represented by (e.g.) Howard and Bannon & Schmidt. As such, it is difficult to assign De Michelis' analysis to the csCW camp. On the other hand, although he applies his framework to categorize IT applications, he is proceeding from a basis of work style; hence it is difficult to see him as purely in the CScw vein. There is something in De Michelis' categorization which seems to capture elements of both the social and the technical dimensions of IT-supported collective activity. What is the basis for his taxonomy being able to bridge these dimensions?
Interestingly, the basis cannot be the obvious candidate -- De Michelis' trinary classification scheme for "cooperation". If one looks carefully at the categories he delimits, the boundaries among them immediately blur, if not disappear entirely. For example, De Michelis himself implies co-decision is a variant of collaboration, qualified with respect to the task being decision making. Coordination can be re-interpreted as either (1) a form of collaboration within which the goal is resource allocation and/or synchronization; or (2) a form of co-decision with the same focus. Finally, it is no great leap to characterize collaboration as entailing significant elements of coordination (e.g., synchronization within a task activity) and co-decision (e.g., selecting what to do next).
If these categories are only fluidly distinguishable from each other, this trifold framework cannot serve as an abstract analytical tool. The question remains as to why it seems to have relative merit as an illustrative device. The real point is not some sort of absolute taxonomic accuracy -- it is De Michelis' shift of definitional emphasis from a general notion of "cooperative work" to more specific, functionally delineable classes of activities. While one may dispute the notion that coordination, collaboration, and co-decision are fundamental categories, it should be clear that they address those activities ascribed to CSCW without falling prey to the ambiguities and subjective values which have plagued attempts to define "cooperative work" generally. In other words, De Michelis managed to enhance the clarity with which these activities are addressable.
This enhancement was obtained by adding discriminatory criteria of specific prospective outputs from the three categories of cooperation. The prospective output of coordination concerns the plans for accomplishing a given task; the prospective output of collaboration concerns the actions by which that task is accomplished; and the prospective output of co-decision concerns policies with regard to some task or topic. These descriptions are more concrete than those emphasizing mutuality of intentions or political ideals like workplace democracy. This concreteness derives from De Michelis' starting from specific forms of work activity (as opposed to some general "cooperative work") and these work activities being utilized in terms of goals or results. De Michelis' primary contribution is therefore the addition of these goal-directed criteria, which enables him to address activities more precisely than (e.g.) Sørgaard or Schmidt.
The foregoing points illustrate that trying to define CSCW via team work qualities leads only to ambiguity. If a CSCW application is defined by work setting, and "all human activity is in some sense 'cooperative' " (Howard, 1987, p. 175), it is difficult to see how any systems could avoid being so categorized. If one speaks broadly enough about the "task" of an entire organization (corporation, agency, etc.), then one can subsume all workers within a group whose goal is achievement of this task, diluting the idea to near-uselessness. The example of De Michelis illustrates that analytical 'traction' can only be obtained when the activity of interest is defined jointly in terms of interacting collaborators and a collective context (e.g., shared goal(s)), not in terms of any a priori quality defined socially or politically.
This, of course, does not close the issue of what we mean by "CSCW". It only demonstrates that the concept, though universally acknowledged as important, is still extremely fluid. Let's now proceed to the application-oriented CScw side of this area -- groupware -- to see if things are any more lucid from the technical perspective.
Groupware is explicitly designed to support collective activity among workers (Dyson, 1989), so it is defined partially by a pattern of usage distributed among multiple collaborators. Even this qualification fails to precisely delineate the field, because the variety of support systems and strategies is as great as the variety of activities and interactions in which people collectively engage (Johansen 1989a). As a result, many diverse artifacts are called groupware, with different authors' categories differentially denoting variants and composites within the potpourri. This variety weakens the label's utility in categorizing software applications (Opper, 1988). Conversely, attempts to define CSCW via enumeration of products or product classes (cf. Lyytinen, 1988; Wilson, 1988; Coleman & Khanna, 1995) have been descriptively useful but ultimately unconvincing with respect to firmly delineating what groupware might be generally.
Grudin (1991, p. 6) outlines unresolved contradictory views on delineating "groupware" vis a vis foundational technologies (database file servers, email, etc.). After dismissing the utility of strict positions on this, he lists (p. 9) three conditions underlying research and development into group IT support: affordable computation; technological infrastructure conducive to communication and coordination; and widespread familiarity with IT tools. It should be clear that these factors are not necessarily linked to the desktop workstations and local area networks commonly cited as groupware foundations. After all, Engelbart's original support tools for "knowledge workers" (1988a; 1988b) -- the harbingers of CSCW -- were conceived and implemented in the days of mainframes and dumb terminals. Had Grudin's three factors been realized through proliferation of 1960's-style centralized IT installations, we might well have seen the groupware phenomenon anyway.
By the same token, the strong association of groupware products with multi-user environments should not be considered an equivalence relation. In a multi-user environment, many people work simultaneously, but there is no connotation their tasks are interdependently executed (at least from the workday perspective of the actors themselves) (Teknikstöd för Grupparbete, 1990). Analogies drawn too strictly from prior multi-user systems have been blamed for contributing to groupware failures (Grudin, 1988). Grudin (1991, p. 9) puts his finger on the key discriminant when he emphasizes that "groupware is designed with the explicit intention of supporting groups, all or most of whose members will use it directly." Ellis, Gibbs & Rein (1993, p. 10) reinforce this orientation when they note the manner in which pre-groupware applications (specifically email and databases) either overlooked or deliberately stymied close coupling among team members.
To summarize, specific systems or environments only provide a context which groups may or may not exploit. There is no reason other than historical coincidence to strongly link the notion of groupware with the desktop microcomputers and LANs stereotypically associated with it. Neither is there any basis for ascribing groupware to IT installations serving multiple people, for populations of end users only provide a context in which task-delineated teams may or may not be realized. Collaborative activity itself is the necessary condition for the label "group", and once this activity is prioritized, it may be supported with multiple styles of deploying the supporting "wares."
As was the case for CSCW, the deployment of groupware applications need not (and in some case did not) map precisely onto (e.g.) conventional operating unit topography. For example, Grudin's (1988) analysis of groupware failures leads him to emphasize focus on operational (as opposed to institutional) entities. His analyses of four groupware types consistently cite conflicts between managers' and workers' interests within organizations, with CSCW failures commonly resulting from managers' decisions to implement systems which better their lot at the expense of others. Grudin (1988) suggests addressing these conflicts by restricting consideration to "...smaller or more homogenous groups", claiming "...there may be less bias when only peer-peer communication is involved than when the communication moves vertically through the organizational hierarchy." (p. 87). Markus and Connolly (1990) extend Grudin's discussion by reference to interdependence among users of specific applications, rather than classes or subunits defined with respect to the organizational map. In both cases, discussion occurs against a backdrop of organizational structure, but analysis defaults to task-oriented relations among interactors. In other words, the relations delimiting the groups of interest are relations of concerted activity, not relations of organizational membership or ranking. This is consistent with the Tofflers' (1990) vision of emerging "flex-firms" and "power mosaics", a prospect in which institutional architectures and rigid intra-organizational relations uniformly and/or universally mappable onto those architectures lose their cartographic and navigational utility.
Because groupware implies collaboration, its delineation and implementation are necessarily intertwined with issues of communication. This is reflected in the typical targeting of groupware products for LAN environments (Johansen, 1989) and telecommunications companies' high profile in CSCW literature and the CSCW conferences to date (Grudin, 1991). At the extreme, Wright (1990) addresses CSCW (as distributed group work -- i.e., different time and/or place) as one of several emerging trends within telecommunications itself.
The need for an emphasis on communications has long been demonstrated by trends toward distributed work through computer networks (e.g., Hiltz, 1984; Masuda, 1982). This already-dated claim was a result of the much older recognition that computers, by virtue of their ability to hold and accelerate information, were communications media (McLuhan, 1964). Beginning with the 1980's, increasingly powerful individual IT tools have been inserted into the workplace. Since the late 1980's, we have been undergoing a "unification" phase in which emphasis is placed on interconnecting those tools. There will necessarily be human and social impacts as this new connectivity is realized; we must anticipate and consider as many of those impacts as we can (Bannon, 1986). Historically, computer hardware development has always progressed faster than software development, while the ambitions of software developers have generally extended beyond their ability to understand work environments. The factors limiting useful applications are those of understanding task domains and applying thoughtful workspace design.
Human/social factors impacts have already been identified in computer-mediated communications systems, e.g., impoverishment of interpersonal orientational cues (e.g., social parameters) in relation to data volume (Hiltz & Turoff, 1981; 1985). In face-to-face interaction, these components are realized using a variety of non-verbal cues. Humans rely on such cues for (e.g.) mutual attention, conversational control, feedback, illustrations, emblematic references, and attitudinal influence. By filtering out such cues, computer-mediated channels reduce the interactional bandwidth and induce a "keyhole" effect. In the absence of contextual or procedural aids, conversants are left to deal with such socializing factors in an ad hoc fashion. This means that there remains a potential for both loss of informative content and introduction of unintended effects.
Although the potential for significant effects from computer mediation of communication was noted many years ago, actual studies on the consequences of shifting interaction to computerized channels have been less prominent. In one such study, Siegel et al. (1986) conducted carefully constructed experiments on group decision tasks enacted through a computer network. They learned (1) there is little constancy of behavior when communications are rerouted through computer networks and (2) behavioral shifts occurred when tasks shifted from face-to-face discussion to computer-based interactions. Restriction or elimination of non-verbal cues from computer-mediated interaction account in part for these results. Such constraints on non-verbal interactional cues was also believed to account for observed effects such as the uninhibited and volatile behavior (flaming) known from the Internet. To give an early concrete example, Shapiro and Anderson (1985) attributed the flaming phenomenon to a set of factors, each of which pertains to the impoverishment of computer-mediated conversational channels and the character of CMC's predominantly text-based interactions.
The two earliest suggested remedies for such problems themselves illustrate the dual technical / collaborative nature which persists in CSCW. The first (technical) trend is to make computer-mediated interactions increase in representational complexity and ease of use, in the hope that such problems will diminish as throughput capacity increases. This is evidenced in the last decade's efforts toward multimedia capacities, non-linear formats (hypertext), and high interactivity. Whether this alone will solve the problem remains to be seen. Enhanced bandwidth and flexibility may not guarantee replicating all the cues used in face-to-face interaction. This caution is reinforced by the fact that we do not have a final understanding of what the essential "cues" or "affordances" may be, given the lack of a comprehensive theory on interpersonal communication.
The second (collaborative) remedy would be to develop, teach, and enforce interactional norms tailored for the new media. Siegel et al. (1986, p. 161) point out "(f)or the purpose of conveying personal and social meaning and for imposing social structure, an etiquette and set of social ethics for electronic communication is still to be developed." Such development has been slow and fragmentary to date. For example, Shapiro and Anderson (1985) outlined normative practices (labelling, emoticons, and encryption) for people using and/or administering computer mediated forums. These suggested measures were already in use at the time of these authors' prescriptions, and they persist to this day -- as does flaming. While this illustrates that use of the new medium encourages development of new etiquette, it does not illustrate that the result need be universally acknowledged. If history is any indication, there is little likelihood that a single such set of norms will prevail. Telephone etiquette varies between organizations and nations, and such variability in an old medium argues for similar variety in the new one. Because existing conversational norms may affect acceptance of any system attempting to impose an interactional style (Hiltz & Johnson, 1989), an interactional stereotype embedded in a software product may not adapt well if transported across organizational or national boundaries (Whitaker, Östberg & Essler, 1989).
When addressing typical CMC tasks like electronic mail and teleconferencing, it is easy to see computer systems as media for the communication of discrete packets of information. This, however, is not a sufficiently rich perspective for CSCW. Addressing the complexities of full-blown work interactivity requires one to dig deeper into the notion of computers as medium -- past the relatively shallow domain of messaging and on to the point where the full range of workspace interaction is exposed. This deeper medium perspective is necessary before we can construct group support systems exhibiting the bandwidth and connectivity to link team collaborators across offices and around the world. Such a general perspective is necessary before we can hope to devise a design and development procedure capable of reflecting those issues omitted or forced into the background by purely technical design orientations -- human factors, authority, autonomy, alienation, and so forth.
These key issues all relate to a given work process -- the missing element in general computer-mediated communication studies. Providing a channel for communications is necessary, but not sufficient, for qualifying an IT artifact as groupware. Meeting room facilities are considered groupware (cf. Johansen, 1989), but much of the supported communication is enacted in the ambient social space, not through the artifacts themselves. More telling is Johansen's (1989) listing of general purpose communication systems in his groupware review, but these are highlighted only within the broad category of "technological support for work group collaboration" -- i.e., foundational artifacts providing a venue for collective activities.
It is access to a common medium within the context of collective activity, rather than communication per se, that is the fundament for groupware. Bannon and Schmidt (1989, p. 364) identify "sharing an information space" as a "core issue for CSCW"; Sørgaard (1988a;1988b) characterizes all IT support in terms of shared material, and De Michelis (1990) cites "information sharing" as the key support for collaborative activity. This shared information space concept is implicit in Engelbart (cf. 1988a; 1988b) and (termed shared environment) is an integral part of Ellis et al.'s (1991, p. 40) definition for "groupware". Both Robinson (1991) and Bannon (1991b) allude to the work of Thompson (1984) with regard to this concept, but do not make any claim that Thompson is the origin of the phrase in the sense that it has become an important "CSCW specific concept" (Robinson, 1991). In any case, these authors concur in a focus on common access to task-specific information rather than the communication links via which that access is realized. This distinction between the technical foundation (communications infrastructure) and the operational benefit (shared information space) is explicit in Bannon (1989) and Grudin's comments on databases versus CSCW (1991, p. 5).
To summarize, groupware systems are by definition channels for some form(s) of communicative interactivity. However, communications systems are groupware only to the extent they lend support to a specific collaborative activity. This specificity is obtained through providing either (a) connectivity for interpersonal co-orientation peculiar to the given collaborative task, or (b) a venue for accretion and sharing of communicative content (e.g., data) peculiar to that task.
To the extent that there has been any analytical "traction" obtained on the issue of what is or is not groupware, it can be summarily illustrated with respect to two examples. One is the canonical "time / space" taxonomy formulated by DeSanctis & Gallupe (1987), subsequently adapted and popularized by Robert Johansen (e.g., 1989b). The second is the groupware overview given by Ellis, Gibbs & Rein (1991).
Johansen's groupware taxonomy uses distribution in time and space to delimit application categories. As he acknowledges (Johansen et al. (1991)), this taxonomy derived from DeSanctis and Gallupe's (1987) illustrative model for categorizing group decision support systems (GDSS). Johansen used this schema to categorize some of his 17 identified types of group support applications (Johansen, 1988; 1989a) into subsets based on the dispersion or copresence of collaborating parties in time and/or space. An abbreviated version of his categorization scheme is illustrated in the table below.
SAME TIME DIFFERENT TIMES
_______________________________________________________
FACE-TO-FACE ADMINISTRATION /
MEETINGS DATA MANAGEMENT
SAME
PLACE
Copyboards Shared files
PC projectors Shift work
Meeting rooms Team Rooms
_______________________________________________________
REMOTE RELIANCE ON
MEETINGS COORDINATION
DIFFERENT
PLACES
Conference calls Electronic mail
Data sharing Forms management
Video/Tele-conferencing Voice mail
_______________________________________________________
This categorization scheme has proven particularly useful both as a means of classifying groupware products and as an illustrative device for demonstrating the types of work environments addressable with such products. Newcomers to the notion of CSCW can easily grasp the time / space permutations and his mapping of product classes onto them. Johansen's time/space diagram has become so ubiquitous that it is now considered something of a CSCW cliche.
Ellis, Gibbs & Rein (1991/1993), declaring "(t)here is no rigid dividing line between systems that are considered groupware and those that are not" (1992, p. 11), conclude it is most reasonable to think in terms of a 'spectrum' within which applications can be classified as groupware. This eminently wise disclaimer is backed up with a pair of dimensions along which such a spectrum may be taken to run. My interpretation of these two dimensions is as follows:
- Commonality of Task addresses the degree to which users of a given groupware artifact jointly participate in a delineable goal-oriented activity. Ellis, Gibbs and Rein associate a low degree of task commonality with "...not working in a tightly coupled mode on a common task..." (1993, p. 10) and a high degree with "...focusing on the same specific task at the same time, and who are closely interacting." (Ibid.)
- Degree of Shared Environment addresses the degree to which collaborators mutually engage within a joint venue for interactivity. Ellis, Gibbs and Rein delineate this dimension with the extent to which groupware artifacts "...provide information about the participants, the current state of the project, and the social atmosphere..." (1993, p. 11)
These authors do not proceed to apply these dimensions in their subsequent discussion of groupware -- instead they fall back to Johansen's time / space matrix and a differentiation by classes of functionality. Nonetheless, their introduction of these two discriminant dimensions provides an improved focus on what constitutes groupware and how the class of groupware is differentiated.
By making time and space the key dimensions for his matrix, Johansen has managed to avoid both the thorny issues of what one means by collaborative work and the non-informative nature of a simple listing of products. The result is a perspective which is useful without either forcing a priori decisions on ill-defined or vague characteristics of the work or limiting oneself to rigid characteristics or classes of IT support tools. This usefulness results from the clarity afforded by focusing on the straightforward dimensions of time and space. This clarity derives in turn from the fact that these key dimensions are coherent discriminants for the class of information or knowledge work activities associated with CSCW. This coherence is a function of the fact that these dimensions provide the basis for mapping possible arrangements among the collaborators and thereby delineating the potential modes of their collaboration.
Ellis, Gibbs and Rein obtain a similarly improved 'focus' on groupware through their two dimensions of task commonality and environment sharing. This improvement results from the clarity afforded by focusing on degrees of specifiable (if not strictly quantifiable) mutuality in activity and operational context. As was the case with Johansen, these authors obtained improved analytical "traction" by keying their descriptions to arrangements among the collaborators.
Early attempts to define CSCW as an area of interest have fallen either into the CScw or the csCW foci. The writers who have managed to somehow avoid the problems of this dichotomy -- Johansen, De Michelis, and Ellis, Gibbs & Rein -- have done so by shifting focus. In the cases of Johansen and Ellis et al. (who refined the CScw angle), referential focus is shifted to a matrix of time and space parameters. In the case of De Michelis (who refined the csCW angle), referential focus is shifted to variations among task goals and/or results. In each case, the dichotomy is circumvented by ceasing to view CSCW (and groupware) as definable with respect to some vaguely defined "collective activity" (or some vaguely related "collective tools") and instead viewing identifiable collective activity as being defined with respect to something more tangible -- an identifiable group who (among other things) share a task and the means for accomplishing that task.
In the case of Johansen, this is illustrated by reference to parameters which map relations among people who are distributed in time and space. This 'human distribution' sets the context for categorizing the various modes of IT support, not the other way around. Ellis, Gibbs and Rein accomplish much the same thing with respect to degree of mutuality regarding task and operational milieu. De Michelis (1990) also prioritizes the collaborators rather than some particular character of the work itself, stating whatever activities we categorize as some sort of cooperative work are known to be thus categorizable by reference to discernible groups and the "...different types of objectives, communication, and relationships..." (p. 2) holding among their members.
Analyzing how the above-cited authors obtained some measure of taxonomic 'leverage' on CSCW and groupware leads to the conclusion that the key criterion for discriminating this ephemeral CSCW area is the group itself -- neither some feature or quality of its activities nor some specific character of the tool(s) it employs (Teknikstöd för Grupparbete, 1990). Dimensions or features pertinent to this criterion include:
This emphasis on the group (or workplace social system) is not an exclusive alternative to (e.g.) time/space parameters or goal specifications; indeed, all three of these aspects mutually influence each other. One can offer examples of groups which do not operate in a collaborative fashion (e.g., the set of all users in the multi-user systems noted earlier), and one may point to software tools designed for individual users which are nonetheless employed for collaborative ends (e.g., word processors used during small group brainstorming sessions). However, there is no such thing as collaboration without there being a group thereby co-defined. Delineation of a "group" is therefore of primary importance in delineating "group work". Admittedly, a number of factors jointly define those workspaces which have heretofore been addressed by CSCW researchers. The point is that those specifiable elements (e.g., time, space, goals, results, and patterns of interactivity) which can be used to 'anchor' such a workspace definition intersect through the collaborating actors themselves.
The shortcomings of CSCW characterizations critiqued earlier derived from assuming the artifacts or work activity exhibiting specific styles (e.g., mutuality of goals; equality of empowerment) were sufficiently determinative to circumscribe the research field and / or the subject phenomenon. Johansen's, Ellis et al.'s, and De Michelis' critical improvements concern criteria which are themselves defined relative to participants in a given collective activity. Time and space intervene between workers, not just their tools. Mutuality of task and environment are delineated among collaborators participating in the former and engaging via the latter. Goals and results are determined, attempted, and evaluated by workers, not some objective guarantor. In other words, groupware is identifiable only with respect to groups, and group work is identifiable only with respect to interactivity.
These points suggest that the most effective vantage for CSCW analysis will be one in which the group is addressed as a collective whose social and functional characteristics are delineated by (and intertwined with) that activity or process within which they participate as such a collective. The 'users' comprise, and thereby delineate, this sense of 'group' by virtue of their reciprocal and mutual activities. This means that the group of interest need not correspond to any discrete region of an enterprise's topology (e.g., an organizational chart). Similarly, the 'activities' comprise, and thereby delineate, this sense of 'group' by virtue of that process (or project, or goal set, etc.) within which specific 'users' participate, either singly or in concert.
This somewhat functionalistic delineation is necessary to match the tools to the praxis which defines the user audience of interest. Terry Winograd noted that simply "...saying that a system is used by a group is not saying much about it at all. It was true of most of the early mainframe applications, and is true of a population of unrelated people using a shared file server." (1992, p. 69) He goes on to explain that the key insight is to frame groupware considerations with respect to a perspective focusing on the group in terms of its collaborative activity. In other words, "...groupware is structured around the way people get their work done. The focus is on designing for the structure of the group activity, rather than the structure of the underlying computer system." (1992, p. 70)
This orientation necessarily argues for a 'user-oriented' or 'user-centered' approach to design and deployment of group IT support. This does not derive from either an engineer's CScw ideal of maximally precise data on the target audience or an action researcher's csCW ideal of maximally 'democratic' engagement on the part of that target audience. Instead, this necessary 'user-orientation' results from the fact that it is that target audience which defines the very domain of intervention (the 'group' or 'team').
The discussion to this point does not (yet) finalize a focus for approaching IT innovations for cognosis and copraxis. Just as there are numerous ways of making a two-dimensional slice through a three-dimensional sphere, there are numerous types of coherent, yet distinct, 'referential slices' one can make through the contextual complex that is a 'group'. It is therefore not enough to appreciate the 'group' as a focal point for enquiry. One must also appreciate and take into account the multifarious ways in which a given 'group' (in the sense outlined above) can be addressed as an object of enquiry and targeted as an object of intervention. Once this necessarily user-centered design process has opened up the possibilities for constructive IT interventions, one must also take into account the diverse ways in which the artifacts realizing that intervention can be viewed and configured. Two examples illustrate this unavoidable attention to perspectival vantage. The first (Malone, 1985) primarily concerns attention to the organizational environment into which IT is to intervene, and the second (Ellis, Gibbs & Rein, 1991/1993) primarily concerns attention to aspects of the IT artifacts themselves.
One example illustrating such differential perspectives is Malone's (1985) discussion of organizational interfaces. Generally speaking, an organizational interface is a projection of the notion of a single user interface (i.e., between an individual 'user' and a single IT artifact) onto the realm of group use (i.e., the interface between a collective and its composite IT support system(s)). Malone defines such organizational interfaces as "the parts of a computer system that connect human users to each other and to the capabilities provided by computers." (1985, p. 66) As a result, any computing system being used by a group should be developed with attention to the organizational interface. The point is that the context for addressing 'group' IT issues is distinct from the prior convention of framing requirements with respect to single 'users', each engaging a single workstation.
This shift of referential framing draws attention to the broader domain(s) of issues entailed in devising IT support for collectives. To exploit this shift to better inform the matching of IT to the character and needs of organizations, Malone claims there is a need for correspondingly novel design theories. He suggests four perspectives from which one may operate in devising such theories:
Malone's recontextualization of IT 'fit' issues with respect to collectives highlights the need for a focus on 'groups' as discussed earlier. His illustrative examples of the four perspectives demonstrates that such a 'group-orientation' is not a complete solution, in that the fundamental delimitations of a 'group', its task(s), goals, and criteria for constructive change must be qualified with respect to one or many candidate orientations. Phrased another way, Malone's treatment of organizational interfaces is both an example of the necessary shift toward 'groups' and a warning of the complexities inherent therein.
Ellis, Gibbs and Rein (1991/1993) note five major 'perspectives' which influence the construction of groupware artifacts, framed with primary respect to disciplinary boundaries. It strikes me that they are secondarily relevant as possible "angles" from which one may consider types or styles of group IT interventions. The authors note that extant groupware applications can be seen as deriving from different permutations of two such perspectives. Beyond that, it is a straightforward step to consider that a groupware artifact might be configured with some "blend" of attributes deriving from these perspectives. The perspectives are:
These authors' recontextualization of groupware implementation issues with respect to different IT specializations demonstrates that the aforementioned 'group-orientation' is not a final solution to groupware implementation. The fundamental delimitations of 'groupware', its functionalities(s), its goals, and criteria for its usage can be qualified with respect to one or many candidate orientations. Phrased another way, Ellis, Gibbs and Rein's treatment of technological perspectives is a warning of the complexities inherent within the notion of providing "support" to groups.
This line of discussion has, in one way, made some progress toward sorting out the dichotomies and dialectics entailed in these things we mean by 'CSCW' and 'groupware.' In another way, however, it has come full circle. Malone's suggested perspectival categories still manifest a division of emphasis between CScw and csCW. His information processing and/or economic perspectives -- the two viewpoints based on quantum data, artifacts, and/or resources -- are clearly tilted in the direction previously labelled CScw. In contrast, Malone's motivational and/or political perspectives -- based on characteristics of work activity and social factors -- would seem to clearly fall under the rubric of csCW.
To a lesser degree, Ellis, Gibbs and Rein's allusions to technical perspectives (as reinterpreted here) can also be seen as falling to one or another side of the CScw / csCW divide. The distributed systems and artificial intelligence perspectives clearly tilt toward CScw, and the communications and social theory perspectives arguably tilt toward csCW, with the human-computer interaction perspective falling somewhere in between (depending on one's precise interpretation or circumstances).
The discussion has highlighted this field's conflictual dualisms, but it has not provided a totalizing synthesis among them. No such synthesis is likely anytime soon. Still, we have refined our outlook on the conundrums. As my Swedish friends might put it, "We are still confused, but on a higher level."
The confusions lingering from the above discussion of CSCW as a subject of research and groupware as a class of IT products have the effect of leaving CSCW (the disciplinary field) similarly ambiguous. There is a widespread feeling that CSCW is not isomorphic with other disciplines or research fields. One could claim this to be simply a matter of CSCW not being isomorphic with any single such field. After all, a certain interdisciplinary flexibility is implied (and demanded) in addressing the confluence of IT and group activities. Still, there is a sense among those folks who consider themselves CSCW researchers that theirs is a field distinguished by something more unique than eclectic confluence. Bearing this in mind, let's now consider how CSCW (whatever it may be) relates to other, more well-recognized, areas of research and development.
The first step in this direction is to survey those fields which are associated with CSCW. The list, unfortunately, is long. Grudin (1991) identifies the MIS and HCI communities as "...the major contributors to groupware development and CSCW research." (p. 12). Ellis, Gibbs & Rein (1991) list "distributed systems, communications, human-computer interaction, artificial intelligence (AI), and social theory" as "five key disciplines or perspectives for successful groupware" (p. 44). It is not likely that the number of candidate 'sibling fields' will be pared down anytime soon, given that it seems every discipline wants to claim a scholarly stake in the 'Information Revolution'.
Such a careful approach is critical, owing to confusions about where CSCW's boundaries are properly drawn. These confusions (delineated with respect to disciplinary 'turf') fall into two cases, which shall be discussed in turn.
This results not only in impractical abstraction and rapid obsolescence, but (more importantly) a blindness to group support issues such as interaction (Bannon, 1991a). This blindness is not overcome through simple analogies between functional unities. Work teams are flexible, dynamic, often transient social networks presumably best analyzed via the social sciences. HF/HCI (as a scientific/engineering enterprise) leaps from the level of the individual to that of a composite, indivisible unit (company, union, etc.) The complexities of work group interactivity cannot be addressed via extrapolations from the individual user to the organization itself or extrapolations from individual users to entire categories or classes of workers (e.g., Östberg & Chapman, 1988, on the "labor force").
Moving to practice, it is much easier to distinguish PD from CSCW. Participatory design is not necessarily 'groupware design'. Although PD is often pursued with groups of workers, there is no presumption the system being designed will be deployed as groupware. PD concentrates on collaboration between designers and users during system development, not on collaboration among users after system implementation (cf. Ehn, 1988b, p. 142). One illustration is the most widely-publicized of all PD exercises -- the UTOPIA project (Bødker et al., 1987), which can be characterized as a sociopolitical exercise in worker empowerment (i.e., PD) focusing specifically on individual interface and functionality concerns (i.e., HCI). On the other hand, participatory design is not necessarily a venue for groupware employment. Much of the PD activity characterized by Kyng (1991) as "an instance of cooperative work" is accomplished without direct IT support of any kind (cf. Ehn & Kyng, 1991). Kyng (1991, p. 72) describes ongoing attempts at Aarhus University to devise "CS" for their "CW". Interestingly, the criteria he mentions as guiding this development are all functionalistic and geared to individual users rather than groups -- e.g., graphic representations and direct manipulation interfaces for better accessibility and usability. This would place that work more within the purview of HCI.
These examples are intended only to illustrate there are as many points of divergence as there are points of correspondence between CSCW and the other fields with which it exchanges ideas and results. HCI has informed us on some aspects of CScw, but it is conceptually and methodologically ill-equipped to address csCW (cf. Bannon, 1991a). PD has addressed csCW, but it is not intrinsically linked to groupware -- either in terms of its tools or its products (Whitaker, Essler & Östberg, 1991). As we have seen, CSCW is not meaningfully reducible to either a purely CScw or csCW perspective. As a result, both these sibling fields are useful, but neither is comprehensive with respect to the phenomena of interest.
In both its senses (object of study and area of studies) the label CSCW is unavoidably intertwined with collective activities and collectively-employed artifacts. However, in neither sense is CSCW comprehensively characterized by exlusive reference to either. Analysis from the basis of stereotyped group activities (i.e., csCW) leads to prescriptions for reconciling sociopolitical factors (e.g., empowerment) when intervening into workplaces portrayed as microsocieties (e.g., PD). This orientation cannot be expanded to encompass CSCW, because it is not necessarily informative on technological issues. Analysis from the basis of work tools (i.e., CScw) leads to prescriptions for reconciling human factors (e.g., performance metrics) when intervening into workplaces portrayed as functional units (e.g., HF/HCI). This orientation cannot be expanded to encompass CSCW, because it is not necessarily informative on social issues. This is not to say the specific fields mentioned (PD and HF/HCI) are of no use to CSCW researchers. A cursory review of these cognate fields' "perspectives" reveals that their relative deficiencies (for subsuming CSCW) complement each other:
These approaches' "blind spots' intersect to profile a "gap" which CSCW researchers should seek to explore. The common link between the csCW and the CScw "perspectives" is the individual human. However, in the former perspective the individual is often subsumed within a bloc (i.e., linked into the workplace social system only in terms of a depersonalized social / political role), while in the latter she is often regarded only in the context of direct engagement with an IT system (i.e., linked into the workplace technical system only in terms of a depersonalized functional role).
The reciprocal influences of social and technical factors require that analysis and intervention be sensitive to both of what have been treated as disparate sides of a csCW and CScw dichotomy. The issue is how to move toward an appreciation of the reciprocities operant in what appears to be better described as a csCW / CScw dialectic. I'm not optimistic about the notion of accomplishing this by somehow summing up the necessarily disparate views of each side taken in isolation. The above analysis of the complementary blindnesses exhibited by (extreme) CScw and csCW suggest progress will entail enquiry targeting the dimensions of intersection and disjunction holding among individual actors and the groups they comprise.
One immediate direction for outlining how to go about this is to extrapolate from the current states of social and technical approaches toward their respective 'inverses' -- i.e., toward novel scopes or foci which address (and hopefully redress) their respective shortcomings. Two specific examples of such 'inverse' agendas are:
Attention to these heretofore undervalued foci will set the stage for better specifying which (and how) IT artifacts can constructively intervene in a work domain delineated with respect to collective action. The csCW and CScw perspectives have traditionally addressed work environments as either unary microsocieties or collections of atomic functional units. Group work is not comprehensively addressable via either of these simplistic vantages, because they obscure the equally valid consideration of work environments as networks of social individuals and / or unary functional systems.
Consideration of group-oriented IT interventions should prioritize identification and evaluation of the target work group. This is a projective conclusion based on the improvement in explanatory power obtained by De Michelis, Johansen, and Ellis et al. as a result of focusing upon specifiable elements (e.g., time, space, goals, results, etc.) pertaining to a team, its members, and the patterns of interactivity among them. Once such consideration has been accomplished, Malone's organizational interface construct provides an example for circumscribing ways of approaching the social / organizational milieu into which IT is to intervene. Similarly, Ellis, Gibbs & Rein's disciplinary perspectives can be usefully taken as an example for circumscribing ways of manipulating the IT interventions. For both the purposes of analysis and intervention, the noted sets of perspectives should be borne in mind -- not as hard and fast 'cookbook' elements, but rather as examples of the diverse vantages from which innovators can and should frame group IT support problem(s) and solution(s).
Baecker, R., Readings in Groupware and Computer-Supported Cooperative Work, San Mateo CA: Morgan Kauffmann, 1993.
Bannon, L., Computer-Mediated Communication, in Norman, D., and S. Draper (eds.), User Centered System Design, Hillsdale NJ: Lawrence Erlbaum Associates, 1986, pp. 433-454.
Bannon, L., and K. Schmidt, CSCW: Four Characters in Search of a Context, in Proceedings of the First European Conference on Computer-Supported Cooperative Work (EC-CSCW '89), London, 1989, pp. 358-372.
Bansler, J., Systems development research in Scandinavia: Three theoretical schools, Scandinavian Journal of Information Systems, Vol. 1, no. 1 (August 1989), pp. 3-20.
Bødker, S., Through the Interface -- A Human Activity Approach to User Interface Design, Aarhus Denmark: Aarhus University Computer Science Department Report DAIMI PB-224, April 1987.
Bødker, S., Knudsen, J. L., Kyng, M., Ehn, P., and K. H. Madsen, Computer Support for Cooperative Design, in CSCW 88: Proceedings of the Conference on Computer-Supported Cooperative Work, Portland, Oregon, September 26-29, 1988, New York: ACM, 1988, pp. 377-394.
Bødker, S., and K. Grønboek, Cooperative Prototyping Experiments: Users and Designers Envision a Dental Case Record System, in Proceedings of the First European Conference on Computer-Supported Cooperative Work (EC-CSCW '89), London, 1989, pp. 343-357.
Bullen, C., and J. Bennett [1990a], Groupware in Practice: An Interpretation of Work Experience, Cambridge MA: MIT Center for Information Systems Research Report no. 205, March 1990.
Bullen, C., and J. Bennett [1990b], Learning from User Experience with Groupware, CSCW '90: Proceedings of the Conference on Computer-Supported Cooperative Work, Los Angeles, ACM, 1990, pp. 291-302.
Cashman, P., and A. Holt, A communication-oriented approach to structuring the software maintenance environment, ACM SIGSOFT, Software Engineering Notes, 5:1, January 1980.
Coleman, David, and Raman Khanna, Groupware : Technologies and Applications, Englewood Cliffs NJ: Prentice Hall, 1995.
Cook, P., C. Ellis, M. Graf, G. Rein, and T. Smith, Project Nick: Meetings Augmentation and Analysis, ACM Transactions on Office Information Systems, Vol. 5, no. 2 (April 1987), pp. 132-146.
De Cindio, F., G. De Michelis, and C. Simone, Computer-Based Tools in the Language/Action Perspective, in Speth, R. (ed.), EUTECO '88: Research in Networks and Distributed Applications, Amsterdam: Elsevier Science Publishers (North-Holland), 1988, pp. 243-258.
De Michelis, G., Computer Support for Cooperative Work, Butler Cox Foundation Report, London, October 1990.
Dennis, A., J. George, L. Jessup, J. Nunamaker, and D. Vogel, Information technology to support electronic meetings, MIS Quarterly, Vol. 12, no. 4 (December 1988), pp. 591-619.
Dennis, A., C. Tyran, D. Vogel, & J. Nunamaker, An evaluation of electronic meeting systems to support strategic management, in Proceedings of the ICIS, 1990, pp. 37-52.
DeSanctis, G., and B. Gallupe, Group decision support systems: a new frontier, DATA BASE, Winter 1985, pp. 3-10.
DeSanctis, G., and B. Gallupe, A foundation for the study of group decision support systems, Management Science, Vol. 33 (1987), no. 5, pp. 589-609.
Dyson, E., Groupware, Whole Earth Review, no. 64 (Fall 1989), pp. 105-107.
Dyson, E., A framework for groupware, in Coleman, David (ed.), Groupware '92, San Mateo CA: Morgan Kaufmann, 1992, pp. 10-20.
Ehn, P., and M. Kyng, Cardboard computers: Mocking-it-up or hands-on the future, in Greenbaum, J., and M. Kyng (eds.), Design at Work: Cooperative Design of Computer Systems, Hillsdale NJ: Lawrence Erlbaum Associates, 1991, pp. 169-196.
Ellis, C., S. Gibbs & G. Rein [1991]. Groupware: Some Issues and Experiences, Communications of the ACM, Vol. 34, no. 1 (January 1991), pp. 39-58. NOTE: This article was reprinted as Ellis, Gibbs & Rein [1993].
Ellis, C., S. Gibbs & G. Rein [1992]. Groupware: Some Issues and Experiences, in Baecker, R., Readings in Groupware and Computer-Supported Cooperative Work, San Mateo CA: Morgan Kauffmann, 1993, pp. 9-28. NOTE: This article originally appeared as Ellis,Gibbs &Rein [1991].
Engelbart, D., [1988a] A Conceptual Framework for the Augmentation of Man's Intellect, in Greif, I. (ed.), 1988, pp. 35-65.
Engelbart, D., [1988b] Toward High-Performance Knowledge Workers, in Greif, I. (ed.), 1988, pp. 67-78.
Engelbart, D., Knowledge-domain interoperability and an open hyperdocument system, in CSCW'90: Proceedings of the Conference on Computer-Supported Cooperative Work, Los Angeles: ACM, 1990, pp. 143-156.
Engelbart, D., and H. Lehtman, Working together, BYTE, December 1988, pp. 245-252.
Gerson, E., and S. Star, Analyzing due process in the workplace, ACM Transactions on Office Information Systems, Vol. 4, no. 3 (July 1986), pp. 257-270.
Grantham, C., and R. Carasik, The Phenomenology of Computer Supported Cooperative Work, Berkeley CA: Interpersonal Software, 1988.
Greenberg, Saul (ed.), Computer-Supported Cooperative Work and Groupware (Computers and People Series), New York: Academic Press, 1991.
Greif, I. (ed.), Computer-Supported Cooperative Work: A Book of Readings, San Mateo CA: Morgan Kaufmann, 1988.
Groupware Resources Guide, Menlo Park CA: Institute for the Future / Groupware Users' Project, 1990 (Second Year Results).
Grudin, J., Why CSCW applications fail: Problems in the design and evaluation of organizational interfaces, in CSCW 88: Proceedings of the Conference on Computer-Supported Cooperative Work, New York: ACM, 1988, pp. 85-93.
Grudin, J., Interface, in CSCW'90: Proceedings of the Conference on Computer-Supported Cooperative Work, Los Angeles: ACM, 1990, pp. 269-278.
Grudin, J., Groupware and CSCW: Why Now? in Hendriks, P. (ed.), Groupware'91: The Potential of Team and Organisational Computing, Utrecht: Software Engineering Reseach Centre, 1991, pp. 3-14.
Habermas, J., Knowledge and Human Interests, English translation by J. Shapiro, Boston: Beacon Press, 1971.
Herbst, P., Socio-technical Design, London: Tavistock, 1972.
Hewitt, C., Offices are open systems, ACM Transactions on Office Information Systems, Vol. 4, no. 3 (July 1986), pp. 271-287.
Hiltz, S. R. Online Communities: A Case Study of the Office of the Future, Norwood NJ: Ablex, 1984.
Hiltz, S. R., and K. Johnson, Measuring Acceptance of Computer-Mediated Communications Systems, Journal of the American Society for Information Science, Vol. 40 (1989), no. 6, pp. 386-397.
Hiltz, S. R., and M. Turoff, The Evolution of User Behavior in a Computerized Conferencing System, Communications of the ACM, Vol. 24, no. 11 (November 1981), pp. 739-751.
Hiltz, S. R., and M. Turoff, Structuring Computer-Mediated Communications Systems to Avoid Information Overload, Communications of the ACM, Vol. 28, no. 7 (July 1985), pp. 680-689.
Hiltz, S. R., and M. Turoff, Virtual Meetings: Computer Conferencing and Distributed Group Support, Chapter 4 in Bostrom, R. P., Watson, R. T., and s. T. Kinney (eds.), Computer Augmented Teamwork: A Guided Tour, New York: Van Nostrand Reinhold, 1992, pp. 67-85.
Hiltz, S. R., and M. Turoff, The Network Nation: Human Communication via Computer (Revised edition), Cambridge MA: MIT Press, 1993.
Howard, R., Systems design and social responsibility: The political implications of 'computer-supported cooperative work', Office: Technology and People, 3, 1987, pp. 175-187.
Johansen, R., Groupware: Computer Support for Business Teams, New York: The Free Press, 1988.
Johansen, R., User Approaches to Computer-Supported Teams, in Olson, M. (ed.), 1989, pp. 1-32.
Johansen, R., presentation to the Teldok Study Tour on "Groupware", Tucson AZ, April 1991.
Johansen, R., A. Martin, R. Mittman, and P. Saffo, Leading Business Teams, Reading MA: Addison-Wesley, 1991.
Johnson-Lenz, P., & T. Johnson-Lenz, Groupware: The Process and Impacts of Design Choices, in Kerr, E., and S. Hiltz (eds.), Computer-Mediated Communication Systems: Status and Evaluation, New York: Academic Press, 1982.
Kelly, J., A reappraisal of sociotechnical systems theory, Human Relations, Vol. 31 (1978), no. 12, pp. 1069-1099.
Kraemer, K., and J. King, Computer-Based Systems for Cooperative Work and Group Decision Making, ACM Computing Surveys, Vol. 20, No. 2 (June 1988), pp. 115-146.
Kyng, M., Designing for Cooperation: Cooperating in Design, Communications of the ACM, Vol. 34, no. 12 (December 1991), pp. 65-73.
Lyytinen, K., Computer Supported Cooperative Work (CSCW) -- issues and challenges, draft working paper, Department of Computer Science, University of Jyväskylä, Finland, December 29, 1988.
Malone, T., Designing Organizational Interfaces, in Proceedings of the CHI '85 Conference on Human Factors in Computer Systems, San Francisco: ACM, 1985, pp. 66-71.
Marca, David, and Geoffrey Bock, Groupware : Software for Computer-Supported Cooperative Work (IEEE Computer Society Press Tutorial), New York: IEEE Computer Society, 1992.
Markus, M., and T. Connolly, Why CSCW applications fail: Problems in the adoption of interdependent work tools, in CSCW'90: Proceedings of the Conference on Computer-Supported Cooperative Work, Los Angeles: ACM, 1990, pp. 371-380.
Nurminen, M., Different Perspectives: What are They and How can They be Used?, in Docherty, P., K. Fuchs-Kittowski, P. Kolm, and L. Mathiassen (eds.), Systems Design for Human Development and Productivity: Participation and Beyond, Amsterdam: Elsevier North-Holland, 1987, pp. 55-62.
Olerup, A., Socio-Technical Design of Computer-Assisted Work, Oslo: University of Oslo Department of Informatics Research Report no. 116, 1988.
Olson, M. (ed.), Technological Support for Work Group Collaboration, Hillsdale NJ: Lawrence Erlbaum Associates, 1989.
Opper, S., A Groupware Toolbox, BYTE, Vol. 13, no. 13 (December 1988), pp. 275-282.
Robinson, M., Double level languages & co-operative working, COSMOS Information Exchange Network, Issue 6 (November 1989), pp. 42-84.
Robinson, M., Computer-supported co-operative work: cases and concepts, in Hendriks, P. (ed.), Groupware'91: The Potential of Team and Organisational Computing, Utrecht: Software Engineering Research Centre, 1991, pp. 59-76.
Robinson, M., and L. Bannon, Questioning representations, in Bannon, L., M. Robinson, and K. Schmidt (eds.), ECSCW'91: Proceedings of the Second European Conference on Computer-Supported Cooperative Work, Dordrecht: Kluwer, 1991, pp. 219-233.
Rouse, W., Group-computer interaction, in Proceedings of the 1973 International Conference on Cybernetics and Society, New York: IEEE, 1973, pp. 145-146.
Schmidt, K., and L. Bannon, CSCW, or what's in a name?, draft working paper, Roskilde: Risø National Laboratory / Frederiksberg: Copenhagen Business School, July 1991. NOTE: This is an extension of Bannon & Schmidt (1989).
Sørgaard, P., A Cooperative Work Perspective on Use and Development of Computer Artifacts, Aarhus (Denmark): Aarhus University Computer Science Department Report DAIMI PB-234, November 1987.
Sørgaard, P., (1988a) Object Oriented Programming and Computerised Shared Material, Aarhus (Denmark): Aarhus University Computer Science Department Report DAIMI PB-247, May 1988.
Sørgaard, P., (1988b) A Framework for Computer Supported Cooperative Work, Aarhus (Denmark): Aarhus University Computer Science Department Report DAIMI PB-253, May 1988.
Taylor, F., The Principles of Scientific Management, New York: W. W. Norton & Co., 1947.
Teknikstöd för Grupparbete, Report of the working group on 'Technology Support for Group Work', Stockholm: UtvecklingsRådet, May 1990.
Thompson, G., Information technology: A question of perception, Telesis, Vol. 11 (1984), no. 2, pp. 2-7.
Toffler, A., and H. Toffler, The Third Wave, New York: Bantam, 1980.
Toffler, A., and H. Toffler, Power Shift, New York: Bantam, 1990.
Trist, E., A socio-technical critique of scientific management, in Edge, D., and J. Wolfe (eds.), Meaning and Control: Essays in Social Aspects of Science and Technology, London: Tavistock, 1973.
Trist, E., The socio-technical perspective: the evolution of socio-technical systems as a conceptual framework and as an action research program, in van de Ven, A., and W. Joyce (eds.), Perspectives in Organization Design and Behavior, London: Wiley, 1981, pp. 32-47.
Trist, E., and K. Bamforth, Some social and psychological consequences of the long-wall method of coal-getting, Human Relations, Vol. 4 (1951).
Whitaker, R., O. Östberg, and U. Essler, Communications and the Coordination of Concerted Activity, Human Interface News & Report, Vol. 4 (1989), pp. 325-338.
Wilson, P. Key research in computer supported cooperative work (CSCW), in Speth R. (ed.), Research into Networks and Distributed Applications, Amsterdam: Elsevier, 1988
Winograd, T. [1987a], A Language/action perspective on the design of cooperative work, Human-Computer Interaction, Vol. 3, no. 1 (1987), pp. 3-30.
Winograd, T. [1987b], Thinking machines: Can there Be? Are we?, Stanford CA: Stanford University Department of Computer Science Report STAN-CS-87-1161 (also available as Report CSLI-87-100), 1987.
Winograd, T., Where the action is, BYTE, Vol. 13, no. 13 (December 1988), pp. 256A-258.
Winograd, T., Groupware and the emergence of business technology, in Coleman, David (ed.), Groupware '92, San Mateo CA: Morgan Kaufmann, 1992, pp. 69-72.
Winograd, T., and F. Flores, Understanding Computers and Cognition: A New Foundation for Design, Norwood NJ: Ablex, 1986.
Wohl, A. (ed.), Groupware -- OA revisited & OA revitalized, The Wohl Report on End-User Computing, Vol. 5, no. 3 (March 1989), pp. 1-6.
Wright, K., The road to the global village, Scientific American, March 1990, pp. 57-66.
