An empirical validation of coupling and cohesion metrics as testability indicators

Abstract

The measurement of the quality attributes of software is a management approach to building quality software product. Testability is a measure of the capability of the software to be subjected to testing. It is a desirable quality attribute of software product because it ensures the reliability of software. This paper presents an empirical validation of coupling and cohesion metrics as indicators of the testability quality attributes of Object-Oriented (OO) software from High-Level Design (HLD) perspective. Open-source OO software samples are used for the empirical analysis and three test case complexity metrics are taken as the measure of the complexity of testing the software sample. The results of the empirical validation showed that the coupling metrics and some of the cohesion metrics investigated are good indicators of the testability of OO software design.