An object-oriented software complexity metric for cohesion

Abstract

The quality of object oriented (OO) software is often measured with the help of OO metrics—A specific type of software metrics that particularly evaluates the features of OO programming such as modularity, understandability, readability, reusability and extensibility. OO metrics that are derived by the concepts of cohesion and coupling depict good measures of module design in such way the module that shares high quotient of data within the module is better than sharing data between them. Thus, the quantification of data/information sharing within the software module is still in the thrust area of research for the approval of module design. The adaptability of existing cohesion metrics is a major concern for the developer as the metric value does not define a clear boundary between low, medium and high cohesions. To overcome this issue, in this chapter, a novel cohesion metric called Cohesion in Method (CohM) is proposed. The metric clearly defines a boundary between the levels of cohesionsby taking the benefits of Jaccard similarity measure. Thus, helps to evaluate software for reusability and maintainability. The results show that the metric comparatively gives better results than the traditional cohesion metrics by clearly differentiating medium and high level cohesions in software module.