Report on a Panel Discussion: Producing Requirements, UKAIS April, 1999

Subject matter (from the panel proposal):

Requirements engineering addresses processes of articulating, abstracting, codifying and managing requirements. In the course of these processes numerous translations occur: tacit knowledge of, say, work practice, must be translated into knowledge that is manifest and can be shared; individual needs and preferences are translated into a social preference order; there is the double translation from natural language to the semi-formal language of say, use cases or Data Flow Diagrams and then to the formal language of computers.

None of this is decidable in advance. Requirements Engineering, as any other IS activity, sits in an organisational context and the precise outcomes are continually shaped and produced by everyday organisational life. The engineer will not understand the marketeer and specify untold bells and whistles. The finance department will gain control of the project at the justification stage and assert its power over the purchasing side by championing a different set of requirements. X will collude with Y to knife Z, using the requirements as a new arena for long-standing conflicts.

The technology context matters too. Changing technologies will help shape choices and established requirements present a solid front to incoming ones. And occasionally an apparently simple new requirement will tip the whole system onto a different, possibly catastrophic, plane. We are after all looking at the year 2000.

Nevertheless RE is a discipline rooted in practice and systems get built. Even if we do not know beforehand where the translations will take us, we do need to understand better and if possible improve the processes involved in the production and translation of requirements.

The panel offered three perspectives:

• Marina Jirotka presented the results of work activities and indicated how they can inform requirements for information systems.

• Richard Vidgen explored interpretations of requirements production using actor-network theory.

• Anthony Finkelstein examined the question of how engineering discipline can be squared with contextually aware practice.

Barbara Farbey chaired the lively discussion which followed, commenting on some of the issues raised with reference to current concerns in industry.

.