Behavior-driven design pattern recovery

In this paper, we present an approach for enhancing program understanding and reusability through a behavior-driven design pattern recovery process. In this context, incorporating behavioral features would characterize the approach as a goal-driven and scalable pattern recovery process. The proposed technique consists of a feature-oriented dynamic analysis and a two-phase design pattern detection process. The dynamic analysis operates on the system's scenario-driven execution traces and produces a mapping between features and their implementation at class level. For the two-phase design pattern detection, we employ approximate matching and structural matching algorithms to identify the instances of the target design pattern that is described using our proposed Pattern Description Language (PDL). The correspondence between system features and identified design pattern instances can facilitate the construction of more reusable and configurable software components to be maintained. We have implemented an Eclipse plug-in toolkit and have conducted experimentation on three versions of JHotDraw systems to evaluate our approach.