New feature adaptation of Github Discussions
Source: Shinshu University

By average number of discussions per day (log scale). Point size represents the number of Members in a project. Credit: Hideaki Hata, Shinshu University

How is a newly added feature on the website GitHub, which is a website essential for many software developers used and utilized? A study led by Shinshu University Associate Professor Hideaki Hata performed an exploratory analysis on this new feature, GitHub Discussions, that is still under development. This study analyzed the beta version that was being tried out on a limited number of projects.

The research questions were as follows:

        RQ1 : How have GitHub Discussions been adopted and used?
        RQ2 : What reasons do developers have to adopt GitHub Discussions?
        RQ3 : How do GitHub Discussions relate and compare to other communication channels?

The international team found that GitHub Discussions are useful not only for software development, but also for managing discussions within a project. However, it is not easy to manage a forum well. Associate Professor Hata found the following points mentioned in section 7 of the research paper are important for improvement.

        Set guidelines for participating in discussions.
        Encourage core project members to participate in discussions.
        Prepare for newcomers.

An international team consisting of researchers with different specialties worked together for analysis. Due to the fact the projects were selected based on criteria that were unknown to the researchers, it was difficult to argue the generality of the results. However, the international research team was effective in performing the preliminary study. Associate Professor Hata hopes to reach out to existing GitHub users and potential users who may use GitHub for collaboration in the future so they could benefit from the research findings. He hopes to analyze and develop a mechanism to manage various processes and communication in software development.

The research was published in Empirical Software Engineering.