Performance and Language Compatibility in Software Pattern Detection

Re documentation and design recovery are two important areas of reverse engineering. Detection of recurring organizations of classes and communicating objects, called software patterns, supports this process. Many approaches to detect software patterns which have been published in the past years suffer from the problems of necessity of reference library, performance and language compatibility. This paper presents a model to solve those problems in software pattern detection. The proposed model solves the problem of necessity of reference library by detecting software patterns using formal concept analysis (FCA). The proposed model solves the problem of performance by using the most efficient algorithm CMCG (Concept-Matrix Based Concepts Generation) for the construction of concept lattice, which is the core data structure of FCA. The proposed model solves the problem of language compatibility by using the language independent meta model called MOOSE for taking the input information. The validity of this model was proved in theory and by experiment.