A Drill-Down Approach for Measuring Maintainability at Source Code Element Level

Measuring source code maintainability has always been a challenge for software engineers. To address this problem, a number of metrics-based quality models have been proposed by researchers. Besides expressing source code maintainability in terms of numerical values, these models are also expected to provide explicable results, i.e. to give a detailed list of source code fragments that should be improved in order to reach higher overall quality. In this paper, we propose a general method for drilling down to the root causes of a quality rating. According to our approach, a relative maintainability index can be calculated for each source code element for which metrics are calculated (e.g. methods, classes). The index value expresses the source code element’s contribution to the overall quality rating. We empirically validated the method on the jEdit open source tool, by comparing the results with the opinions of software engineering students. The case study shows that there is a high, 0.68 Spearman’s correlation, which suggests that relative maintainability indices assigned by our method express the subjective feelings of humans fairly well.