Assessing maintainability change over multiple software releases

Abstract

The focus of the paper is to reveal the relationships between software maintainability and other internal software quality attributes. The source code characteristics of five Java-based open-source software products are analyzed using the software measurement tool SoftCalc. The relationships between maintainability and internal quality attributes are identified based on the Pearson product moment correlation analysis. Our results show negative correlations between maintainability and some well-known internal software quality attributes, as well as the ones between maintainability and complexity metrics. Particularly, according to our results, the Number of Data Variables Declared and the Decisional Complexity McClure Metric have the strongest correlations with maintainability. The results of our study, that is to say, knowledge about the relationships between internal software quality attributes and maintainability, can be used as a basis for improvement of software maintainability at earlier stages of the software development process. Copyright © 2007 John Wiley & Sons, Ltd.