I think this falls on the DevOps area (me, @toxinu and @gabor) but we don’t have capacity for these low-priority tasks now.
This can be taken up in any Ocim related projects though.
I don’t have any hard opinions here, but don’t we want our work to have some visibility? As an OSS, being on Github is kind of unavoidable, to be honest. But if we don’t really care about being widely open to contributions I agree that we should migrate to Gitlab.
Well honestly it seems that nobody else is interested in Ocim; it’s too specialized to OVH and/or OpenCraft. However, I think the new version v2 that uses Tutor and supports other providers like DigitalOcean and AWS will be much more useful to others.
So for v1 I don’t think it matters where it’s hosted, and we should do whatever is consistent and easiest for us.
So… which is the canonical hosted source though currently? Github or Gitlab? I feel like it’s somewhat high priority to sort this out, because conversations are getting more and more fragmented between the two hosted copies.
I think this is the kind of feeling that github is wanting to create, and it comes from github’s ubiquity. However, I’m not sure it’s accurate. I’ve seen many many successful OSS projects that aren’t on github. Being on github may lower the barrier to contributions, but maybe this just results in lower quality contributions on average?
Anyway, that’s a discussion for another topic. The concern I have here is that we have two hosted copies of ocim right now, with the master branch at different revisions, and both have active issues and pull requests. We must consolidate that to one, before the situation worsens.
GitLab is open sourced. Meanwhile, GitHub is closed and proprietary. Therefore, most projects created should be on GitLab.
…
[But] when projects contain code that is meant to be merged into one of the projects from the edX GitHub organization, then they should be created on GitHub.