Remote Cooperation:
CSCW Issues for Mobile and Tele-Workers
Alan Dix and Russell Beale (eds.)

As Thimbleby and Pullinger (this volume) suggest "for mobile, personal users of computer systems, CSCW is in a mess". Although it is in a mess, there are other reasons than they suggest.

By way of a preface to the issues raised in this chapter Thimbleby and Pullinger's argument as presented in their paper in this volume is a useful starting point. They suggest that systems designed to support cooperative activity require a great deal more consideration of the individual user's point of view. They also note that many CSCW systems are designed with 'implementation-based' properties in mind: systems are 'over-specified' technically to be as powerful as possible since the problems of CSCW are often attributed as 'technical' ones. This argument - that the success or otherwise of CSCW systems rests somewhere 'outside' of the system and its implementation properties, usually in organisational or group interfaces - is of course is a familiar argument in the literature on CSCW.

However, Thimbleby and Pullinger's argument is more sophisticated. They suggest that 'implementation-based properties' are not those which are the legitimate concern of the user in the context of cooperative work. In the place of such properties can be placed what might be called 'activity-centred requirements' - 'transparency', 'consistency' and 'grace' - and corresponding 'observational properties' of the designed system (such as 'observational consistency', for example). Now the properties which are important in determining the success or otherwise of a CSCW system are those which flow from building systems with these general requirements in mind. One of these, simply put, is that what is important for the user is what the user can observe and act upon: the actions of the various users working cooperatively need to be rendered consistent when looked at from the individual user's point of view.

What this argument does is to place the individual user's concerns squarely in the centre of the design of CSCW systems. This is obviously desirable. However, by focusing the argument on the cooperative element of managing information this tends to miss the point that much of what is important in managing information within a group context is not defined by its cooperative nature, but by its personal nature. A system with the observational property of consistency, for example, will not help users manage 'personal' information - or will do so only weakly; what it may do is help them view certain kinds of data as consistent within a cooperative setting.

An important move is to take an 'inside-out view' of CSCW. This view focuses on 'personal information management' first and foremost as a building block to looking at the ways in which cooperative systems can be designed for the cooperative management of information. Of course, this argument applies equally to the design of single-user applications and individual work - although it is debatable whether such 'non-cooperative' applications, and such work, actually exist (Riddick and Thomas 1992). Some of the issues discussed in this chapter - such as the notion of 'fragile expertise' - are being applied in our work to the design of single-user applications and their interfaces (Thomas 1993, 1995; Thomas and Meech 1994, for example). However, it is a particularly relevant direction to the design of cooperative working systems in which the task of information management has become decentralized and fragmented, particularly in the case of users whose point of entry to a cooperative systems may be a personal computer used outside of a physical, organisational or group context and not permanently connected to the high-bandwidth networks which are usually considered a defining feature of CSCW systems.