Onboarding to Open Source Projects with Good First Issues: A Preliminary Analysis

While open source software (OSS) development projects always seek contributions from newcomers to make the projects sustainable, newcomers often face with a challenge of onboarding. For developers with little experience, it is difficult to decide which of the many OSS projects to contribute to and which issues to tackle. To mitigate the barrier against onboarding, some OSS projects start using a label so called good first issue (GFI) which indicates the issue is easy to resolve and suitable for newcomers to tackle. The final goal of this study is to construct a model for recommending good first issues to help reduce maintainers' manual effort of selecting and labeling issues as GFIs and at the same time to help find GFIs suitable for each newcomer. Toward the final goal, we analyze GFIs in GitHub to deeply understand the current state of the GFI mechanism and its impact on new members' onboarding. This paper reports a result of our preliminary analysis of GFIs, based on 9,475 GFIs collected from 11 famous OSS projects in GitHub. We find that (1) developers who have resolved GFIs have 92 fewer median PR (pull request) posts than developers who have resolved regular issues (i.e., developers tackling GFIs have less experience, compared to other developers), (2) on average, GFIs are resolved 32.5% more than regular issues (i.e., GFIs are easier to resolve. The dataset of GFIs might be able to be used as a training set for recommending GFIs to developers with less experience), and however (3) the percentage of developers who keep contributing to the same project even after resolving GFIs varies greatly from project to project (24.9% to 83.9%) (i.e, GFIs guide newcomers' onboarding only in particular projects). Based on the analysis result, we discuss the contents of GFIs which tend to keep newcomers on OSS projects.