A Cohesion and Coupling Driven Ontology for Measuring Software Quality at Code Levels

The need for quality and reliable software systems has led to the development of rigorous software measures of quality at code level. One of the basic problems in software measurement is that many of the existing measures do not measure what they claim to be measuring. Chidamber and Kemerer metric suites were used to measure cohesion and coupling in six industrial systems. Findings of the study indicated that cohesion and coupling measured quality at code level in four of the systems in terms of the systems being highly cohesive and low in coupling. Cohesive systems 1, 2, 4, 6 had median values [0,1]. In these systems the level of coupling is acceptably low. In terms of correlations cohesion and coupling, cohesion and size, coupling and size all correlated significantly. The study concludes that cohesion and coupling were useful quality software measures in the studied systems.