DPJF - Design Pattern Detection with High Accuracy

摘要

Automated design pattern detection (DPD) is a challenging reengineering task. The detection quality (precision and recall) of DPD tools has so far been insufficient to make DPD integral part of current IDEs and development practices. In this paper, we present a novel approach and a related tool, DPJF. For all implemented pattern detectors and all projects used for evaluation, DPJF achieves 100% precision and the best recall of all evaluated tools. Still, its analysis speed competes with the fastest existing tools. The high detection quality is achieved by a well-balanced combination of structural and behavioural analysis techniques whereas the good performance is achieved by emprically validated simplifications of the individual techniques. Our results lay the basis for routine application of DPD in program comprehension and let DPJF pioneer novel uses of DPD for software quality assesment and improvement.