Goal-Based Domain Modeling as a Basis for Cross-Disciplinary Systems Engineering - Intentional Perspectives on Information Systems Engineering - page 96

ever keep pace with the development of new systems in changing domains”. But
generalized domain knowledge seems to rarely benefit the narrowly focused SMEs
and means to update the domain model according to experiences in concrete projects
have been presented here. Sure et al. [ 41] investigate the use of ontologies for knowl-
edge management in enterprises. Their approach can be considered heavy-weight in
that they expect an explicit ontology engineer to take care of the evolution of the
ontology. SMEs rarely have the manpower (or financial scope) for such a position.
Instead they need a more pragmatic approach that immediately and tightly integrates
with their development practice.
In our approach, design goals, functional requirements, and agents are used
mostly as an organizing principle for information management from the perspective
of design product reuse. Colette Rolland's long-standing research on requirements
process modeling [35] offers an interesting complementary perspective that deserves
further attention in our future research. Her Map formalism interprets goals as inten-
tions that can be achieved within the process by alternative strategies where each
strategy can have steps from different disciplines; this Map information could be
used as an additional aid in the similarity search process presented above, even
though it has yet to be determined whether such a strategic process modeling
approach is really helpful in the SME context where usually only a few people,
who know each other's way of working well, have to cooperate. However, in [ 36] ,
she additionally points out that the Map formalism can also be exploited as process
guidance in the model evolution step to maintain the alignment among the different
submodels.
5 Conclusion
In this chapter, a domain-model based approach to requirements engineering for
project-oriented small and medium-sized enterprises has been presented. This
approach tackles the identified challenges by building on a goal- and agent-oriented
model based approach. The few and simple concepts of i ∗ have proven to be suit-
able to address interdisciplinarity. The project-oriented development is addressed
by a domain model that captures the extensive knowledge of an SME in a particular
field and thereby accelerates requirements capture. A similarity search supports the
identification of reusable solutions in earlier projects. The necessary evolution of
the domain model to maintain its usefulness is considered while still ensuring the
accuracy of similarity search results. Support for deriving subsequent development
artifacts via semi-automatic transformation finally completes the proposed tool set
and contributes towards quality assurance.
The application of the proposed approach has been exemplified for the field
of control systems. While the concrete domain model is of course domain spe-
cific, the basic ideas behind the approach are applicable to other application
fields with similar characteristics. For example, systems for access control and
burglary warnings for buildings offer similar characteristics as control systems
engineering. It is customer- and project-oriented since each security system is