Assessing maintainability: Information theory, metrics and iterative software processes

The purpose of this dissertation is to assess maintainability and evolution of object-oriented (OO) software developed using highly iterative or agile software processes with a focus on the utility of information-theory based metrics in this analysis. To do this, we first explored the utility of existing OO metrics and metric suites to assess maintainability in terms of fault-proneness in these types of systems, using statistical models predicting fault-proneness as a basis for the assessment of these metrics. We conclude that existing OO metrics are useful in measuring maintainability in these systems due primarily to the dynamic change in these systems from one version to the next. We examined the utility of information theory-based metrics to assess maintainability in terms of complexity. Finally, we examined maintainability in terms of design instability and degradation. We developed two new information theory-based metrics and empirically showed them to be useful in measuring design instability and degradation in highly iterative or agile software projects.