The Coherence Project: towards an ethnographically informed process for requirements engineering

Computing Department, Lancaster University, UK

This decade has seen ethnography become increasingly popular in the requirements engineering (RE) community. Studies have been performed in a variety of domains [1, 2, 3]. By spending time alongside workers, observing what they do, ethnographers develop a deep understanding of the work. Ethnographers can therefore provide designers with detailed insights into the work as it is actually performed, presented in the language and terminology of the users. Ethnographic studies can uncover subtle features of the social nature of work that are vital to successful operations, yet at the same time appear to be so trivial that other techniques can miss them. This is in contrast to some other human-centred approaches, which tend towards studying simulations of work in artificial laboratory-based settings, and imposing their own vocabulary to describe the work.

Ethnography has a great deal to offer as a technique for RE, but a number of issues limit its use in practice:

Time. Ethnography can be a very lengthy process, lasting months or even years in the context of social research. Much shorter time-scales are required for the RE process.

Results. Ethnography tends to produce a great deal of detailed, textual description as a result of performing a study. Effectively communicating these findings to requirements engineers is not straightforward.

Culture. There are significant differences in language and culture between sociologists and software developers. This can lead to problems of communication between the two groups.

Abstraction. It is difficult to draw design principles and other abstract lessons from a technique that is concerned with the detail of a particular situation.

Skill. The lack of a systematic approach to conducting ethnography makes the technique dependent on the individual ethnographer’s skill.

We have been addressing these issues in our research for several years [4, 5, 6, 7, 8]. The Coherence project represents a break from previous research by addressing them from a different perspective. Coherence builds upon the previous work at Lancaster on ethnographically informed design, and on viewpoint-oriented requirements engineering [9, 10]. Rather than modifying how ethnography is undertaken or presented to systems design, in Coherence we are looking to modify the requirements process so that it is influenced from an ethnographic perspective. The aim is that, in so doing, we will arrive at an ethnographically informed approach to requirements engineering that can be applied by RE practitioners without necessitating extensive training in sociology. Further, our aim is to utilise existing standard notations from the software engineering community in order to communicate the results of Coherence analysis to the rest of the development process.

In particular, we have explored using built-in extensions to UML (the Unified Modelling Language) to express the results of an ethnographic study using a standard object-oriented notation [11]. We have also demonstrated how the Coherence approach can be used to generate initial Use-Case models and therefore provide a route into a standard development process [12]. By providing interfaces to standard notations and processes such as these, we hope to have created an approach that is more likely to fit with existing work practices of software developers in industry.

A trial of Coherence in a High Street Bank has recently been completed. This talk will concentrate on the issues arising out of the trial, and their implications for Coherence’s development and acceptability as a practical approach to RE, informed by ethnography. They include:

Process. Despite the appearance of a rational, step-by-step process, in practice this just isn’t the case. Should we even be looking for this at the earliest stages of RE? If we do, are we in danger of missing out on important details in order to follow the process?

Time. It is common for work to occur over regular periods–daily, weekly, monthly. You can’t hope to capture all of this in a limited period of observation spent in the organisation. How do you then make sure that important time-related aspects of work are not overlooked?

Details. Ethnography is all about the details. An ethnographically informed method should be able to index into detailed accounts–or audio or video tape–where necessary, without overwhelming developers with detailed textual description as the only means of describing the work.

Focus. How can the method assist the analyst in focusing on relevant (social) aspects of the workplace? This is also a process issue. Initial open-ended observation provides a feel for the workplace, areas for concern, etc. At some stage there is a need to become more focused and concentrate on specific issues. The method should assist in providing the focus, and prompting the analyst with areas of potential interest.

References

Bentley, R., Hughes, J.A., Randall, D., Rodden, T., Sawyer, P., Shapiro, D. and Sommerville, I., Ethnographically-Informed Systems Design for Air Traffic Control. In Proceedings of ACM CSCW'92 Conference on Computer-Supported Cooperative Work (Toronto, Canada, 1992) ACM Press, 123-129.
Blythin, S., Rouncefield, M. and Hughes, J.A., Never mind the ethno stuff–what does all this mean and what do we do now?: Ethnography in the commercial world. Interactions 4, 3 (1997) 38-47.
Heath, C. and Luff, P., Collaboration and control: crisis management and multimedia technology in London Underground control rooms. Computer Supported Cooperative Work 1, 1 (1992) 69-94.
Hughes, J., King, V., Rodden, T. and Andersen, H., Moving out from the control room: ethnography in system design. In Proceedings of the ACM 1994 Conference on Computer Supported Cooperative Work - CSCW’94 (Chapel Hill, NC, 1994) ACM Press, 429-439.
Hughes, J.A., O’Brien, J., Rodden, T. and Rouncefield, M., Designing with Ethnography: A Presentation Framework for Design. In Symposium on Designing Interactive Systems–DIS’97 (Amsterdam, Netherlands, 1997) ACM Press, 147-159.
Hughes, J.A., O’Brien, J., Rodden, T., Rouncefield, M. and Sommerville, I., Presenting ethnography in the requirements process. In Proceedings of RE’95 (York, UK, 1995) IEEE Computer Society Press, 27-34.
Sommerville, I., Rodden, T., Sawyer, P. and Bentley, R., Sociologists can be surprisingly useful in interactive systems design. In People and Computers VII: Proceedings of the HCI’92 conference (York, 1992) Cambridge University Press, 341-353.
Sommerville, I., Rodden, T., Sawyer, P., Bentley, R. and Twidale, M., Integrating ethnography into the requirements engineering process. In Proceedings of RE’93 (San Diego, CA, 1993) IEEE Computer Society Press, 165-173.
Sommerville, I. and Sawyer, P., Viewpoints: principles, problems and a practical approach to requirements engineering. Annals of Software Engineering 3, (1997) 101-130.
Sommerville, I., Sawyer, P. and Viller, S., Viewpoints for requirements elicitation: a practical approach. In Proceedings of the IEEE International Conference on Requirements Engineering - ICRE’98 (Colorado Springs, Colorado, 1998) IEEE Computer Society Press, 74-81.
Viller, S. and Sommerville, I., Coherence: an approach to representing ethnographic analyses in systems design. Human—Computer Interaction 14, Special issue on representations in interactive systems development (1999)
Viller, S. and Sommerville, I., Social analysis in the requirements engineering process: from ethnography to method. In Proceedings of RE’99 (Limerick, Ireland, 1999) IEEE Computer Society Press,