CSCW systems have been developed in a number of contexts, normally following one of a number of communications models. They often require extensive use of expensive hardware and software, as well as making very heavy use of network resources. It has been shown that extremely effective collaboration can be instituted with only the most basic facilities. Diaper and Beer [1], for example, considered the situation where journal reviewers could pool comments for evaluation by the Editorial Board, and onward transmission to the authors. It was also shown that it was unnecessary to insist on highly sophisticated and specialist software for effective collaboration. This is important when large numbers of users, such as referees, are only intermittently involved, and the activity is only a small part of their normal work. A large investment in equipment resources or their time in learning a large and complex system is not justified by the anticipated usage.

This study considers the interaction between the group of project visitors for the British Computer Society Awards Scheme [2] and uses as its base the BSCW system [3,4]. BSCW was chosen because all application specific code is managed on the BSCW server. The clients required to take part in the collaborative working only require a basic World Wide Web navigator in a standard configuration.

For CSCW systems to become fully accepted, packages need to be developed that are of such general utility and easy configurability that they can be readily adapted to meet the needs of each specific application. It was only for example, when the highly flexible nature of relational database systems became fully appreciated that they became all pervasive in commercial applications.

It is through studies such as the one described in this paper that CSCW tools can be refined to meet the requirements of a wide range of realistic applications, other than those originally considered by their authors. In this way, tools and techniques can be developed to reduce the amount of work required to build such applications. It was to meet this objective that the initial user analysis described in this paper was performed using basic Object Oriented techniques [5], supported by a standard CASE toolset (in our case Select-IT [6]). This provided an excellent framework on which to build an effective system.

Initial analysis was based on Use Case Analysis [5]. The principal Actors and processes were identified, as shown in Figure 1. This proved to be an extremely useful activity, as it concentrated further development on those activities necessary for the management of the Awards assessment process, rather than all the other activities associated with the Awards, such as the running of presentation events. These are primarily internal to the British Computer Society, and only marginally involve the Visitors and Referees who are expected to be the primary users of this system.

It was always the intention to manage the BCS Awards Workspace on its own BSCW server. This allowed customization to be undertaken without regard to other workspaces and which the contents are to be used. The requirement is to associate capabilities to each of these roles. Participants can then be invited on the basis of the role that they are expected to undertake, and the relevant capabilities can be associated with them when they register.

The original plan of encouraging all project documentation and reports to be entered directly into the system was soon found to be over ambitious. The documents concerned often describe material of commercial significance. This means that not only must the security measures be watertight, but also every entrant and referee must have full confidence in it. This is very difficult to achieve with a completely untried system. A two-stage implementation scheme has therefore been adopted. In the first year, the system described is being used to distribute the majority of papers and reports for each project. They are entered by the BCS Administrator, after scanning if necessary. This considerably reduces the amount of paper being distributed to Visiting Panel Members, without increasing the Administrator’s workload significantly. This is because the process of scanning and uploading the documents requires less effort than the photocopying and mailing of sets of documents to each panel member.

The use of the Meetings facility in BSCW to assist in the arrangement of Visits was very much appreciated by Visitors. Even though the facility is used rather differently than originally envisaged by the designers of BSCW. Even though the Administrator still arranges the Visit by other means, and confirms by letter, there is a single repository of papers associated with that Visit. Travel arrangements are much more easily reconciled within the visiting team, as they all have access to the same, hopefully current, information.

The Visitors also appreciate the collection of all the Visits with which they are involved in their own personal Meetings folder. This is arranged in date order, and provides a convenient reference point for checking on modifications to arrangements during a very busy period.

The management of a workspace of the size and complexity of that described in this paper requires a significant level of management and support. One of the major outcomes of this project is to identify the areas in which this support must be targeted. When using a Web based front end, as is provided with BSCW, it is very easy to assume that there is already widespread acceptance of such tools, and that little if any training and support is necessary.

Particular difficulties were experienced with the initial registration process as the boiler plate text of the electronic mail message sent when participants are invited to undertake a role was found to be extremely off-putting. It is just about acceptable when the participant has initiated the registration process, and an Email message is expected, simply to give instructions as to how to complete the registration process. As an invitation to a referee, or to a project leader whose project has just been nominated, it is completely unacceptable. The standard invitation screen allows additional text to be added to the standard electronic mail message. This was found to be totally inadequate when large numbers of new participants needed to be introduced to the system. Each electronic mail message had to be individually customized, rather than simply including the participant’s details and project information, as the current system allowed, by using carefully constructed letter templates.

Issues such as this need considerably more consideration as CSCW become more widespread, and are used amongst much wider and more diverse communities.

The customization undertaken in the project was very specific. We believe that many of the facilities included have general utility that, if combined with appropriate analysis methodologies, could ease the implementation of such systems. If CSCW systems are to be of more general utility they have to be made very much easier.

The use of Use Case Analysis techniques rapidly identified the activities involved in the part of the Awards process to be supported in this project, and the roles of the various participants. This helped considerably in identifying the needs of individual users, and determining the best ways of supporting them.

Implementation plans are also important to acceptability. This paper looks at an application area that has a clear annual operational cycle. An incremental approach delays the full benefit gain until a later cycle. It does however, provide a high level of flexibility in incorporating improvements and additional features as development progresses. Participants rapidly become accustomed to using the existing features, and are soon clamoring for more. It was fortunate that a major requirement, the wish to reduce the amount of paper being distributed through the post could be met very easily by simply building a hierarchy of workspaces within BSCW. The meeting facility also already existed, and with very little customization, could be adapted to provide a much-needed aid to the arrangement of Visits. As more information is available directly from the Web, the ability to post pointers to the relevant pages was very much appreciated.

BSCW has proved to be a very flexible platform on which to build this application, and has allowed the development of a flexible collaborative working environment that can be used effectively by a significant population of participants to undertake the roles that they have agreed to perform. The Awards process is essentially a voluntary effort, and the system described could not be implemented without the full agreement and consent of all involved.

The author acknowledges the considerable help given by both fellow members of the Visiting Panel and members of the permanent staff of the British Computer Society. This work would also have been impossible without the help of project leaders and referees who have assisted in the design and evaluation process.

[1] Diaper, D. and Beer M. D. (1995) 'Collaborative Document Annotation using Electronic Mail', Computer Support of Collaborative Work (CSCW) 3: pp297-325, 1995.

[3] Bentley, R., Applet, W., Busbach, U., Hinrichts, E., Kerr, D., Sikkel, K., Trevor, J. and Woetzel, G. (1997), Basic Support for Cooperative Work on the World Wide Web, International Journal of Human-Computer Interaction, Special Edition on novel applications of the World Wide Web.

[4] Bentley, R., Horstmann, T., Sikkel, K. and Trevor, J, (1995), Supporting Information Sharing with the World Wide Web: The BSCW Shared Workspace System, 4^th International WWW Conference, Boston, December 1995, pp63-74.

[5] Eriksson, H. and Penker, M., (1997) UML Toolkit, Wiley.

[6] Alan, P. and Frost, S. (1997) Service-Based Architectures – Unlocking the Potential of Component Based Architectures, American Programmer, July 1997.

[8] Coularis, G. and Dollimore, J. (1994), A Security Model for Collaborative Work, Technical Report 674, Dept. of Computer Science, Queen Mary and Westfield College, University of London.

[9] Coularis, G. and Dollimore, J. (1994), Requirements for security in cooperative work: two case studies, Technical Report 671, Dept. of Computer Science, Queen Mary and Westfield College, University of London.

[10] Sikkel, K. (1997), A Group Based Authorisation Model for Cooperative Systems, Proceedings of ECSCW’97, Lancaster, UK, September 1997, Kluwer.

[11] Shen, H. and Dewan, P. (1992), Access Control for Collaborative Environments, ACM Conference on Computer-Supported Cooperative Work (CSCW’92), Toronto, Canada, pp51-58.

[12] Leveridge, P. (1997), Campus World and BT’s on-line educational services, BT Technical Journal, 15, 2, April 1997, pp 126-131.

[13] Sandhu, R. S., Coyne, E. J., Feinstein, H. L. and Yourman, C. E. (1996), Role-Based Access Control Models, IEEE Computer, February 1996, pp38-47.