Collaborative Requirements Engineering: Bridging the Gulfs Between Worlds - Intentional Perspectives on Information Systems Engineering - page 367

improved to fit different ways of working. Even though the storyboards, prototypes
and scenarios were used in separate analysis sessions with academic researchers
and the PCT analysts, they formed a common ground understanding of alternative
views which evolved into specification of core common functionality of the sys-
tem, with customised versions for each stakeholder group to meet their needs and
values.
7 Conclusion
The contributions of this chapter have been to develop Rolland's vision of
method engineering from the perspective of discourse theory and natural language.
Application of Clark's common ground as a framework for critiquing represen-
tations and techniques throws light on their relative contributions to one of the
fundamental problems in RE: how to reconcile abstract and concrete views in sys-
tem development, so that a mutual understanding of requirements, the software
design and domain constraints emerges. Combination of representations has also
been researched in Rolland's method engineering [ 26, 27] to suggest how scenarios
and models in particular can improve understanding of requirements. Other exam-
ples of using scenarios and concrete representations to challenge models and goals
can be found in obstacles analysis for KAOS modelling [ 37] and scenario-based
validation in ScenIC [ 18] .
Use of scenarios, prototypes, and design rationale formed the kernel of the
SCRAM method [ 30] ; however, experience demonstrated that too many represen-
tations can overload users, resulting in poor focus on key issues [ 33] , in spite
of the use of design rationale and scenario walkthrough to structure sessions. So
while juxtaposing representations is advantageous, optimal use still poses several
research problems. This raises a fundamental question in the method engineering
debate: whether it is best to provide users and requirements engineers with a tool-
kit of representations and techniques and let them take the decisions on how to use
them during the process; or the alternative of trying to provide prescriptive cook-
book guidance for different situations. Rolland tended towards the flexible tool-kit
approach [ 19, 26, 28] , with which I agree. However, I argue that a theoretical frame-
work, such as the common ground described in this chapter, can give researchers and
practitioners useful criteria to compose the RE process and choose representations
to suit their particular circumstances.
References
1. Baddeley AD (1986) Working memory. Oxford University, Oxford
2. Beck K (1999) Extreme programming explained: embracing change. Addison-Wesley,
New York
3. Brennan SE, Clark HH (1996) Conceptual pacts and lexical choice of conversation. J Exp
Psychol: Learning, Memory Cognition 22(6):1482-1493