Let’s talk about it – Jason Tsay

We know that discussions on pull requests are important, but what we do not know yet is how these discussions impact projects. To study this, Jason has studied pull requests. He randomly selected over 600,000 PRs from 12,483 projects and then manually selected a number of them to study.

On these PRs, he used Grounded Theory to categorize them. What he found was although pull requests are obviously very technical, many of the comments are social and political:

  • There clearly is ‘newcomer politeness’: core always thanks newcomers and apologizes for delays in commenting
  • Sometimes alternative solutions are suggested, to maintain integrity
  • Also, sometimes code developers refuse pull requests because the issue is not deemed very relevant (feature outside of scope of the project)
  • Finally, sometimes outsider or third party developers try to influence the outcome of the discussion, for example if they want to have support for their own particular library or framework.

In summary, different roles have different interests:

2014-11-18 13.51.09Update: there is a preprint available here: http://www.jsntsay.com/publications/fse2014-letstalkaboutit.pdf