Developer involvement considered harmful?: an empirical examination of Android bug resolution times

Abstract

In large scale software development ecosystems, there is a common perception that higher developer involvement leads to faster resolution of bugs. This is based on conjectures around more "eyeballs" making bugs "shallow" -whose validity and applicability are not without dispute. In this paper, we posit that the level of developer attention as well as its extent of diversity influence how quickly bugs get resolved. We report results from a study of 1,000+ Android bugs. We find statistically significant evidence that attention and diversity have contrasting relationships with the resolution time of bugs, even after controlling for factors such as interest, importance, dependency etc. Our results can offer helpful insights on team dynamics and project governance.