Goal-Based Domain Modeling as a Basis for Cross-Disciplinary Systems Engineering - Intentional Perspectives on Information Systems Engineering - page 83

Managing the fitness relationship between business needs and system functionality
is becoming a continuous task. Rolland [ 36] proposes an explicit management of
the fitness relationship as a formal model mapping between business models and
system models.
As we are moving into the era of embedded software-intensive systems, the map-
ping between business goals and software technologies needs to be augmented with
a cross-disciplinary mapping within the systems themselves, in particular relating
hardware systems (with aspects from mechanical and electrical engineering) with
the software systems that control and monitor them. Such tasks are by no means
limited to large companies e.g. in the utilities, automotive, or aeronautical fields. In
fact, much of the innovation comes from small and medium engineering enterprises
(SMEs) who either sell products directly to end customers or supply the big players
with innovative designs and components based on inter-disciplinary development
projects. The research question addressed in this chapter is therefore how we can
extend the idea of fitness to such project-driven cross-disciplinary SMEs.
Important success factors for project-driven SMEs are their flexibility, innova-
tiveness, and customer orientation [ 4, 11, 21] . For them, requirements engineering
activities are part of the offer development where timing and cost constraints are
very tight. Pre-planned product line engineering [ 33] unfortunately does not work
for these SMEs without abandoning much of their customer-orientation and flexibil-
ity, since they cannot commit to the necessary prerequisite of some domain stability
[ 21] . On the other hand the knowledge an SME has gained throughout previous
projects is its core asset and must be reused extensively during later projects [ 3] .
Accordingly, they need means to keep track and internally make available their
extensive knowledge [ 24] . Eventually, this knowledge must be easily adaptable and
extensible in order to evolve with new innovations and gained experiences.
The situation is complicated further if the business of the SME is interdisci-
plinary. A typical example is control system development [ 1] in automotive design.
More and more car control functionality - such as engine management or driver
assistance systems - is realized in software on electronic control units [ 7] . Only this
way, quality goals such as comfort, safety, and energy efficiency can be tackled. But
the required interplay of theories, methods and tools from control engineering and
software engineering, and the resulting interworking of people with different world
views and concerns, demands more advanced modeling and management methods.
To address these issues, this chapter presents a goal- and dependency focused
domain model based approach to SME-oriented requirements management .The
approach takes agent goals and their interdependencies as a starting point, building
on the i ∗ framework [ 45] , but elaborates them to a fully model-based domain-
specific information management approach. It consists of:

a simple, but suitably expressive common notation based on i ∗

• that allows to

capture the requirements of all disciplines,

• a domain model offering a collection of functional requirements patterns as well

as non-functional goals for a particular (sub-)domain and tailored to a particular
SME,