4 ms·
Github (I presume it's all GH projects) should come up with a system that makes it easier to either take over stewardship of a codebase, or make active forks mu
by Cthulhu_ 5y ago
Github (I presume it's all GH projects) should come up with a system that makes it easier to either take over stewardship of a codebase, or make active forks much more visible - I mean in theory if the steward of one project has effectively abandoned it, someone else can fork it and go from there. But in practice, that fork fixes one issue and gets abandoned as well, and nobody can find it (or nobody actively looks for it).
But if you can make a fork and petition for it to be seen as the "active, maintained" version, it would work a lot better I think.
In theory, open source is all about forking and the like. In practice, it's a lot of islands with single people or organizations at the "top" dominating the project, with any forks having no chance at survival unless they get merged in again.
- tn1 5y agoIn the Perl world (and possibly others), the CPAN maintainers will transfer ownership/maintainership privileges to you if you can prove you haven't been able to contact the original author (typically by sending an email and CC'ing the perl5-modules mailing list). This works well when there is one or two distributions that are the canonical way to solve a task. Obviously not possible for GitHub as a whole, but perhaps we could reimagine how GitHub interacts with specific language/tool/whatever communities.
- franciscop 5y agoThis is very similar to how npm works! I got some simple name packages after the original authors didn't answer like `server` and `files`.
- Jenk 5y agoand is also the double edged-sword that permits bad actors to take over popular packages and inject malware and other nasties :(
- codetrotter 5y ago> But if you can make a fork and petition for it to be seen as the "active, maintained" version, it would work a lot better I think. In practice I fear that this would lead to abuse where scammers (or what you want to call them) fork dead projects but then add crypto miners, adware and ransomware to the code and then they create a bunch of dummy commits to make it seem active to automated systems and to people only taking a glance at the project. And all of that can happen already of course, but if you then allow these people to automatically be assigned status of the canonical active maintained version then much more people will end up installing these malware-infested versions of the software.
- jerf 5y agoI think there's a balance between what they show now and what I'd like them to show, for sure. I certainly do not want GitHub to try to declare what an "official" repository is, in any form whatsoever. GitHub simply is not in a position to make that declaration, philosophically, practically, or any other way. On the other hand, the network view, which is the closest thing we have right now, is pretty hard to read for the question of "where are active forks right now"? I just sampled them now to make sure I'm not shooting my mouth off, and I think a major problem is that the graph doesn't have a strong visual difference between an active fork with many commits and contributors and someone who happened to commit something yesterday against an old (perhaps "official" fork), and there doesn't seem to be any sensible sorting that I can make out, either. Plus the UI on that thing is just funky in general; I think I see lines going off the right of some of these networks entirely but I can't seem to scroll right or left. Upshot is, I'd rather they just figured out how to improve that graph, or generally build a "more active forks available here" that can be obtained through a clickthrough, but not have GitHub try to guess at who's "official".
- akent 5y agoCheck out https://useful-forks.github.io/ https://useful-forks.github.io/ if you haven't seen it, agree it'd be great to have this kind of thing a lot more visible particularly on repos with no commits for like 6+ months or something.
- withinboredom 5y agoJust because a project doesn't have any recent commits, doesn't mean it doesn't work. Often you can look up it's forks and find a more active fork.
- leifg 5y agoI like that! Unfortunately for libraries you will also have to have a system in place for all the library hosting providers (npm, rubygems …) to transfer ownership as well.
- toast0 5y ago> Github (I presume it's all GH projects) should come up with a system that makes it easier to either take over stewardship of a codebase, or make active forks much more visible They've got that weird graph thing of fork activity, assuming people forked on Github, that has worked well enough for me to find a less recently abandoned fork on the things I've run into.