Cooperative work is seamless

CSCW seems to be in a dilemma. On one hand there is a position which aims at the provision of systems that prescribe interaction but which, in doing so, restricts coordinative interaction, and on the other hand a position which aims at radically flexible systems that do not prescribe interaction but which, in doing so, leaves it to the users to cope with the complexity of coordinating their cooperative activities. How do we move beyond this schism? In addressing this problem, the paper discusses two general modalities of articulation work — ad hoc alignment and improvisation on the basis of mutual awareness versus coordination in terms of a predefined flow of work — and argues that these modalities are seamlessly meshed and blended in the course of real world cooperative activities. On the basis of this discussion the paper outlines an approach which may help CSCW research to bridge this gap. Two approaches to the issue of how to combine a prescribed flow of work with flexibility can be distinguished. The first approach attempts to reconcile these two apparently contradictory goals by looking for a language incorporating an ontology of cooperation which in turn is derived from a conceptualization of how cooperating actors construct their cooperative environment. A typical example of this approach is the Language/Action Perspective which led to the development of systems like Action Workflow [37] and, more recently, the Milano System [1]. Irrespective of their differences, both systems provide a set of primitives which is claimed to be sufficiently expressive to cover the needs of users in articulating their activities. Flexibility is thus claimed to be achieved by means of a universal language which provides a set of primitives derived from the ontology. The second approach tries to achieve the same reconciliation of flexibility and prescription by selecting a suitable metaphor for modelling cooperative work. Such metaphors usually privilege a specific type of workflow, and takes other types as subordinated and derived: the flow of tasks, as in the case of Regatta [58], or the flow of documents, as in the case of LinkWorksTM1, or the flow of communication, as in the case of Strudel [48] or Conversation Builder [9; 32]. In this approach flexibility is achieved by providing primitives for manipulating the flow, for example by allowing the flow to be defined dynamically and incrementally. 1. THE SCHISM IN CSCW When considering the many imaginative and inspiring computer-based facilities for supporting cooperative work which have been proposed in recent years, research prototypes as well as commercial products, one can distinguish two competing approaches with respect to the underlying conceptions of ‘computer support’: On one hand, a position which suggests to support cooperative work by providing normative models of cooperation so as to regulate the routine coordinative activities and thereby enable cooperative ensembles to perform more reliably and efficiently, and on the other hand, a position that suggests to support cooperative work by offering a ‘shared work space’ through which actors can interact directly, i.e., by means of generic competencies such as talking, gesturing, pointing, monitoring etc., without other restraints than the constraints of limited bandwith and so on. In the meanwhile, a rich amount of reflections rooted in a bunch of empirical studies have led to considering these approaches problematic. On the one hand, the approach based on the Language/Action Perspective is considered controversial [4; 11; 55] and, more generally, together with other workflow technologies, not informed by empirical studies [10]. On the other hand, the metaphor-based approach seems too narrow and to be causing a fragmentation of the technological support to cooperation. In fact, each system was focused on only one aspect of coordination and could not be integrated with technologies supporting the other aspects.2 An additional milestone was Suchman's contribution which emphasized the situated nature of action [56]. These problems, together with the actual difficulty of progressing with significant steps towards flexibility, generated an increasing interest in looking for an alternative approach. This trend was also motivated by the raising of new cooperative settings where coordination shows the property of an inherently As a natural reaction to the limitations of office automation, the first position was first on the scene and generated a range of different support systems. While these systems are different in many ways it is beyond the scope of this paper to enter into a detailed discussion of these differences. It is sufficient for our purposes to consider how the different modeling approaches deal with the flexibility required by the changing needs of the target work setting. This is one of the basic requirements that became evident as experiences with their use were gained and reported [e.g., 34, ff.]. 1 LinkWorksTM is a groupware product by Digital. 2 In fact, some recent evolution of these proposals strive to overcome this limitation as, for example [e.g., 35].