Is using mercurial+bitbucket instead of git+GitHub limiting my contributors?
I recently found out that our most prolific third party contributor actually uses git to do and test his work, and only converts to mercurial commits for the pull requests.
Am I shooting myself in the foot by using mercurial?
We originally chose mercurial because our users (and thus likely contributors) are scientists (as are we), and many are unfamiliar with software development tools and practices, and frankly, nothing is as newbie friendly as tortoisehg (even as a mercurial expert tortoisehg is my preferred way to do 99% of my work with mercurial, and I will miss it sorely enough to attempt to write a clone of it for git if we switch to git). But it seems maybe the people most likely to contribute are the subset familiar with programming, and they are more likely to use git and github.
tl;dr should I move my project to git and github?
6 comments
[ 4.6 ms ] story [ 27.3 ms ] threadAbout the source code control. generally, I would prefer git. You can take a look at https://www.atlassian.com/blog/git/git-vs-mercurial-why-git for making decisions.
As wolco said, you're surely missing out on network effect, people won't find your project if they're just searching GitHub (as many people, unfortunately, do lean towards).
I would ask your users though, rather than HN. HN is going to trend towards what's popular in general, rather than what your users like to work with. Find out how your users would feel about Git or a Git-based software solution, maybe mirror your project somewhere, be it GitHub or GitLab and let people test out what it'd be like to work from the Git side of things.
Hoping to get new contributors is great, but make sure your existing contributors are on board, or even enthusiastic, before you look for a migration.
[0] https://tortoisehg.bitbucket.io/screenshots.html
Searching probably isn't how people find out about our project, though I'm not sure. We have a published paper that gets citations, so people find out that way, and I suspect it might be mostly word of mouth. But we really don't know.
You're right, we should ask them! But we have zero telemetry so we only know about the users who participate in the mailing list or issue tracker or otherwise contact us.
I would say the people making active participation in your mailing list or issue tracker, and especially those making patches to your actual software, likely represent the most active/invested members of your community, and definitely the best people to solicit opinions from. :)