Analysis on Relationship between Code Quality and Code Coverage in an XP Environment: a Case Study on the SWURV Project

The thesis used hypothesis testing and correlation analysis methods to explore the relationship between structural code coverage and the quality of software developed in an eXtreme Programming (XP) environment, via a case study of a commercial software product. We find that improving code coverage is helpful to detect residual defects, but it is not enough, and we also need other testing, like acceptance testing, in the process of XP software development to provide good quality software products. In addition, in order to investigate why the strength of association between code coverage and residual defect density is not as strong as that presented in prior work, a detailed defect root cause analysis is performed, showing that over 96% of bugs cannot be detected by improving code coverage. Based on the defect categories and distribution of defect root cause, six improvement actions are proposed for future XP projects.