Analysis of Early Aspects in Requirements Goal Models: A Concept-Driven Approach - Transactions on Aspect-Oriented Software Development - page 56

candidate aspects in goal models by using the tasks they crosscut, and RGT
could provide support for doing so.
To investigate construct similarities and differences across individuals using
RGT, we assume that concrete entities, such as tasks, in the given context can
be mutually understood among stakeholders. It might not be the case that peo-
ple can precisely and accurately interpret other people's elements, because the
actual phrasing and labeling of elements will have crucial impact on our pro-
posed method. Our overall perception is that people make good approximations
when trying to understand concrete and pertinent concepts, so our assumption
is “good enough” for applying RGT to deal with early aspects alignment in
goal models. Besides, any comparison of conceptual systems necessarily involves
approximation since a complete conceptual system may involve indefinitely com-
plex relations and different concepts will never be identical in all respects [43].
Finally, like any process, the quality of the output of our RGT-based method
is only as good as the quality of the inputs and the controls over the process.
Our inputs rely on various stakeholders in the problem domain, on which early
aspects research is based. And we believe that our approach not only has strong
yet flexible controls over the extraction, exchange, comparison, and assessment
process, but also has a profound and solid foundation, namely the PCT, of align-
ing concepts in different viewpoints. Our output, such as the clustering result
shown in Fig. 9 and the vocabulary map in Fig. 10, can enable the requirements
analyst to both gain insights into stakeholders' use of domain concepts and to
generate plausible hypotheses to guide further early aspects analysis activities.

4 Early Aspects Analysis

This section describes the application of FCA to supporting trade-off analysis
and conflict detection of early aspects in requirements goal models. We also
discuss the seamless integration of FCA with RGT so as to form a coherent
concept-driven framework. Before diving into the details of FCA, we examine the
connection that “concept merging” (step 3 in Fig. 1) makes between candidate
aspects alignment and analysis.
The early aspects alignment method introduced in Sect. 3 can assist in high-
lighting discrepancies over the terminology being used and the concepts being
modeled. Concept merging in goal models can thus be considered as grid merging
of consolidated tasks and reconciled softgoals. We assume that in the adapted
merged view, there may only be one contribution relationship between a task
and a softgoal: either being negative or being positive, but cannot be both, 2
i.e., the merged model exhibits an internal consistency with respect to soft-
goal contributions. For example, although both the customer and the developer
expressed “Usability” in their original views (Figs. 2 and 3), a terminological

2 We simplify the scale for measuring softgoal contributions to be “positive” and
“negative”. In Fig. 6, “break” and “hurt” are associated with “negative”, “help”
and “make” correspond to “positive”, and “neutral” means there is no contribution
relationship between the task and the softgoal.