Practice has proven that softgoal interdependency graphs (SIGs) are a very useful tool for recording and reasoning about non-functional requirements [4], [6], [8]. Practical reuse and design of SIGs, however, rely on the capacity of manually identifying reusable structures, interdependencies, and eliminating harmful redundancies. In this paper, we tackle this shortcomings by firstly proposing a frame representation, that allows us to represent SIGs in terms of lattice-like generalization/specialization hierarchies that can be queried. Secondly, we introduce a closeness criteria that, in the absence of perfect matches, will allow querying for similar or "good enough" approximations. This mechanism is particularly useful in the identification of softgoal inter-dependencies. Finally, we argue how the proposed strategy can be used to help identify conflicts in SIG graphs.
展开▼
