Static analysis and run-time coupling metrics

The relationships between coupling and external quality factors of object-oriented software[1] have been studied extensively for the past few years. For example, a clear empirical relationship between class-level coupling and the fault-proneness of the classes have been identified by several studies. A number of statistical techniques, principally Agglomerative Hierarchical Clustering (AHC) analysis, Byte Code Instrumentation are used to facilitate the identification of such objects. Dynamic coupling indicates the strength of association established by a connection from one software module to another at runtime. Despite the rich body of research in the field of software measurement, dynamic coupling measurement for Aspect Oriented software is still missing. A dynamic coupling measurement framework for AspectJ[10] programs is presented in this paper. The framework consists of a suite of measures for both method-level and class-level coupling relations. This paper also presents a new approach towards static analysis, in particular class analysis to the computation of dynamic coupling measures and is designed to work on incomplete programs.