It’s been one month since #Guix migrated to Codeberg.
We’ve had 138 code contributors on that month compared to 102 on the previous month according to Git, or +35%.
Of course it’s too small a sample to draw any conclusion but let’s hope it continues that way.
It is not only about contributions, but about the pace maintainers can deal with them, which was one of the main reasons to migrate. The relevant statistics are here
https://codeberg.org/guix/guix/activity
What’s weird to me is the need to approve changes before merging, which requires intervention of two maintainers instead of just one, effectively doubling the requirements of the main bottleneck in Guix and the chances of changes being forgotten.
But overall, I feel like cooperation is smoother and clearer; new people join, quite a few contributors appear to be more productive, and more folks contribute to the tools/infra.
Héhé! CI had been an argument in favor of the migration. 😁
But hey it should be improved soon because more hands and easier to better support, I guess.
$ git shortlog -nse --after=2.month --before=1.month | wc -l
104
$ git shortlog -nse --after=1.month | wc -l
142
?
If yes, I’m not sure this counts any Codeberg effect. Because, for example, these top contributors over the past month:
417 Nicolas Graves <ngraves@ngraves.fr>
270 Liliana Marie Prikler <liliana.prikler@gmail.com>
135 jgart <jgart@dismail.de>
105 Vinicius Monego <monego@posteo.net>
95 Sharlatan Hellseher <sharlatanus@gmail.com>
67 Ricardo Wurmus <rekado@elephly.net>
would have contributed equally; as many other in the list. In other words, most of these contributions had been merged without Codeberg involved. 😁
Moreover, these numbers about code contributors say nothing about how they contributed (still via email guix-patches@gnu.org or via PRs).
Moreover, consider:
417 Nicolas Graves <ngraves@ngraves.fr>
15 Nicolas Graves via Guix-patches via <guix-patches@gnu.org>
is counted twice for that month; among many other examples. And that also happens over the previous month. But it’s hard to tell if the “duplicate author” numbers are the same; a bit more or a bit less then it drastically changes the percentage. 😉
The sample is still too small to say something relevant, IMHO.
My point is:
We cannot say (yet!) something about the positive effect of the migration, as @civodul seems suggesting. 🙃
Today, we can only say: The migration does not have a negative impact on the contributions. And that’s already very nice! 🎉
Somehow, “not a negative effect” doesn’t imply “a positive effect”. It isn’t a bolean logic: “not negative” <=x=> “positive” or “not positive” <=x=> “negative”. Touching my noose 🤡 every morning isn’t negative for my health condition, so do we say then it has a positive effect on my health condition? 🤔
My completely uneducated guess is that moving to a forge will result in a small increase in the rate of submissions, but a bigger increase in the acceptance rate, as it makes reviewing and discussing patches easier
This is a small personal instance of Bonfire in the Fediverse.
