Generalized coupling measure for aspect-oriented systems

Abstract

Coupling is an internal software attribute that can be used to indicate the degree of interdependence among the components of a software system. Coupling in software has been linked to predict external software quality attributes such as fault-proneness, ripple effects, changeability, impact analysis etc. Despite an interesting body of work for measuring coupling in Aspect-Oriented (AO) Systems, there is no complete generic coupling framework for AO systems. In this paper, we have proposed a generic coupling measurement framework that takes into account three, the most well known families of Aspect-Oriented Programming (AOP) languages, AspectJ, CaesarJ and Hyper/J. This paper contributes to an enhanced understanding of coupling in AO systems, which in turn helps to (i) define new coupling metrics which permit the analysis and comparison of Java, AspectJ, CaesarJ and Hyper/J implementations, and (ii) integrate different existing measures and examine same concepts from different perspectives.