The Coupling Between Objects metric (CBO) is a widely-used metric but, in practice, ambiguities in its correct implementation have led to different values being computed by different metric tools and studies. CBO has often been shown to correlate with defect occurrence in software systems, but the use of different calculations is commonly overlooked. This paper investigates the varying interpretations of CBO used by those metrics tools and researchers and defines a set of metrics representing the different computational approaches used. These metrics are calculated for a large-scale Java system and logistic regression used to correlate them with defect data obtained by analysing the system's version tracking records. The different variations of CBO are shown to have significantly different correlations to defects. Regarding results, a clear binary divide was found between CBO values which, on the one hand, predicted a defect and, on the other, those that did not. The results, therefore, show that a clarification or unambiguous re-definition of CBO is both desirable and essential for a general consensus on its use. Moreover, applications of the metric must pay close attention to the actual method of calculation being used and, conclusions and comparisons made as a result.

No abstract

No abstract

In this paper, a tool is described for visualising the Coupling Between Objects (CBO) metric for Java systems, decomposing it into coupling collaborators and using colour to denote the object-oriented mechanisms at work for each couple. The resulting visualisation is also envisaged to be useful for general program comprehension and is integrated into Java development in the Eclipse IDE. Evidence is also given that the visualisation may help detect classes tending to be less fault-prone than would be expected from inspection of their CBO values alone.

No abstract