Friendica Social Network (Leipzig)

"GitHub OSS Governance File Dataset", Yen et al., 2023

❓ How many #github repos have a governance.md?

➡️ ~1,600,000 🐘

❓ Of those, how many have the governance.md in their root directory? (I.e. remove dependencies)

➡️ 1,899 🐭

❓ Of those, how many have have at least one issue/commit? (I.e. 'significant')

➡️ 710 👀

https://arxiv.org/abs/2304.00460# #gov #governance

GitHub OSS Governance File Dataset

Open-source Software (OSS) has become a valuable resource in both industry and academia over the last few decades.

^arXiv.org

Doug Webb 8 Monate her •

Screenshot from paper supporting original post, text reads:

2) Repositories: Based on GitHub search results, there are more than 1.6M projects with a GOVERNANCE.MD file. However, not all of them are signaling meaningful information as these projects often do not have a GOVERNANCE.MD file for their own projects but include other packages or dependencies that have GOVERNANCE.MD files. 1 Therefore, to reduce noise in the data set, we used the REST API’s search code endpoint 2 to fetch only the repositories that contain a file with a GOVERNANCE.MD (case-insensitive and non-exact match) in the root directory. 3 The initial data set consists of 1,899 unique projects that have a GOVERNANCE.MD file in their root directory. To ensure the projects are meaningful, we only keep projects that have at least 1 commit/issue. This results in the final data set of 710 projects. We gathered a list of basic metadata for each repository, such as the repository name and the filename of the governance file.

Doug Webb 8 Monate her •

Authors note limitations

1) github API non-deterministic

2) there are many other places project may keep their governance (e.g. in their readme, any other filename, not on github...)

Screenshot from paper supporting post, text reads:

Limitations We note that this dataset has two major limitations regarding completeness and potential bias. Completeness 1) Because GitHub’s search API doesn’t return a deterministic and full set of results, the dataset is not a complete set of all GitHub-hosted repositories with a governance file. 2) As we only collected projects which contain the GOVERNANCE.MD file in the root directory, some GitHub-hosted projects are missing from our dataset as they might organize and store their governance files differently. For example, some projects put their governance files directly in the readme file.

Doug Webb 8 Monate her •

In other words: a hell of a lot more than 710 github projects have governance documents.

600mb dataset and the scripts they used here: https://zenodo.org/records/7530768

Open-source Software Governance Documentation Dataset on GitHub

This dataset contains 710 GitHub-hosted OSS projects, which contain a governance file in the root directory of the project. It also contains commits, issues, and comments on each project.

^Zenodo

⇧