Exploring the relationship of history characteristics and defect count: an empirical study

During the lifetime of a project, a huge amount of information is generated, e.g. in versioning systems or bug data bases. When analysed appropriately, the knowledge about the previous project characteristics allows estimating the project's future evolution. For example, it is very valuable to know particular history characteristics of a file indicating its fault proneness because it helps testers to focus their testing effort on these specific files. In this paper, we present the results of an empirical study, exploring the relationship between history characteristics of software entities and their defects. For this purpose, we analyze 9 open source java projects. The results show that there are some history characteristics that highly correlate with defects in software, e.g. the number of changes and the number of distinct authors performing changes to a file. The number of co-changed files does not correlate with the defect count.