An Experimental Study Comparing the Lack of Cohesion in Methods (LCOM5) Metric and the Cohesion on Method Attribute (COMA) Metric

Abstract

This paper presents an experimental study on the comparison between the LCOM5 cohesion metric and the COMA metric, The LCOM5 metric is based on the calculation of a number Pairs of roads with no traits in common.

Software metrics can help address the most critical issues in software development and can provide support for planning, predicting, monitoring, controlling, and evaluating the quality of both software products and processes. Quite a number of object oriented cohesion metrics have been proposed; we identified lapses in the definition of some of the object-oriented cohesion metrics. In this paper we propose a new metric for measuring cohesion at the design level. The purpose of the study is to identify and address the defects of the LCOM5 metric The methodology relied on designing a tool using C# that calculates the consistency of the object-oriented program with LCOM5 and COMA.

we can see that LCOM5 is solely based on attribute referencing and thus will not show an increase in the cohesion of a class where there is method invocation in addition to attribute referencing, Unlike COMA, LCOM5 could not differentiate between the cohesion of this class But COMA was able to differentiate between the cohesion of these classes.