Theoretical Evaluation of Software Coupling Metrics

Software module coupling is an important design parameter that can be used in various studies including fault prediction, impact analysis, re-modularization assessment, software vulnerabilities assessment, etc. However, two couplings can vary in important coupling factors' coverage. This paper aims to evaluate the coupling metrics to their coverage of important coupling factors'. We perform a thorough survey of coupling metrics, that ends up with the collection of numerous coupling metrics. After that, we evaluate these metrics by their coverage of important coupling factors. The mapping of coupling metrics with the coupling factors shows that the coupling levels been considered by many metrics. Yet, the difference between these levels has been ignored by most of the metrics. Moreover, the broadness, hiddenness, and rigidness of data flow are ignored by most coupling metrics. Apart from that, the combined effect of these aspects is ignored by all the metrics.