A Metrics Suite for Measuring Class Interaction at Run Time through DCM

The first step taken while developing software, by an analyst is to construct a sequence diagram which describes the interaction that must occur between classes. The sequence diagram does not show the interaction of classes at run time. It shows static coupling, i.e. it only acknowledges the number of interactions between the classes. There have been many studies done regarding the relationship between coupling and external quality factors of object-oriented software. A common way to define and measure coupling is through structural properties and static code analysis. However, because of polymorphism, dynamic binding, and the common presence of unused or dead code in commercial software, the resulting coupling measures are not precise as they do not perfectly reflect the actual coupling taking place among classes dynamically. This paper proposes the use of the estimated frequency of the use cases and propagates these frequencies through the sequence diagram to estimate dynamic coupling. This can be done by using Dynamic Clustering Mechanism (DCM) in which classes which interact with high frequency are clustered or grouped together (called as hot spots), which are highly dynamically coupled. With this evaluation we will be able to determine which hot spot are indeed relevant and actually deserves close attention from the designer w.r.t design of each class.