Semantic-Based Interaction Detection in Aspect-Oriented Scenarios

Interactions between dependent or conflicting aspects are a well-known problem with aspect-oriented development (and related paradigms). These interactions are potentially dangerous and can lead to unexpected or incorrect results when aspects are composed. To date, most aspect interaction detection methods have been based either on purely syntactic comparisons or have relied on heavyweight formal methods. We present a new approach that is based instead on lightweight semantic annotations of aspects. Each aspect is annotated with domain-specific markers and a separate influence model describes how semantic markers from different domains influence each other. Automated analysis can then be used both to highlight semantic aspect conflicts and to trade-off aspects. We apply this technique to early aspects, namely, aspect scenarios, because it is desirable to detect aspect interactions as early in the software lifecycle as possible. We evaluate the technique using an industrial case study and show that the technique detects interactions that cannot be discovered using syntactic techniques.