Before Github, there was no collection of open source repositories that are easily searchable, making it easy to find and promote open source software. I am not aware of any alternative that ever did or does a better job at making open source contributions that easy. Even when I try to use codeberg as an alternative, my Github repos will always get more contributions. No idea how we could even begin to change that.
I don’t agree with the whole list, but the CLA requirement and corpo projects pinky-promising they’d never do a bad thing and then going to do a bad thing as soon as their investors demand returns is certainly a major risk and harm. I’ve started self-hosting everything for my personal use, and if it’s not AGPL, then I assume at some point I’m going to get fucked and shouldn’t rely on it.
Also, the endless stupidity around everyone using Discord as their primary means of communication, discussion, issue reporting and whatnot. Politely, fuck Discord, and fuck anyone who thinks Discord is the right way to make anything accessible to the public.
There’s lots of other alternatives, including ye olde IRC and forums and even simple mailing lists - and no, I don’t mean ‘sign up for our newsletter!’ nonsense, but an actual real mailing list. And, if you want something a little more modern, there’s always Matrix which is probably feature-complete enough to compete with whatever you’d want to use Discord for anyways.
Politely, fuck Discord, and fuck anyone who thinks Discord is the right way to make anything accessible to the public.
You had me at “politely”.
I’ve been finding Zulip quite helpful. It’s threading model is great and they overall focus quite a bit in the project coordination use-case. You can either self-host it or pay for their managed hosting (which is free for open-source projects), and you can add a plugin to make static HTML pages of streams (aka channels) in order to make stuff indexable and searchable (and iirc this is getting polished and built into Zulip’s core).
If you care about accessibility, email is still the best choice — it’s mostly text-focused, doesn’t need an account (besides what is universally seen as the most basic Internet identity), truly decentralized and has mature tooling. I just haven’t found a really good mailing list archive web UI. HyperKitty is good, but isn’t quite there for me. lists.sr.ht is neat, but lacks a lot of features. Above all, indexability and searchability (from inside the UI itself) is key.
Oh that’s nice. Hadn’t seen their stuff before but that looks like a MUCH better option than Matrix, if you want a shiny gui app and that kind of experience. And can’t argue with the pricing if you’re running an open-source project with it, though I suppose you can make a comment that it’s still got a vendor lock-in problem.
And 100% agree that email is the gold standard, still, and yeah, nobody has really come up with an amazing web UI for searching list archives.
Emails are nowhere near being competitive with discord. Sure, they’re technically more accessible, but in practice they are much less usable by random people which in turn will just avoid interacting or contributing with your project.
I’m not sure how they are less usable than Discord. “Everyone” (using quotations here because it’s not an absolute thing, but it’s almost so) knows how to e-mail, it’s one of the most fundamental Internet skills. Using Discord, however, is not, for a large amount of people. Sure, most developers either have had contact with Discord at some point or are capable of figuring it out just fine anyway. But seeing as FLOSS really shouldn’t just be about developers (as Drew points out too) and as end users should also be accounted for, e-mail as a basis for coordination and support is a very valid choice.
It’s pretty much account-less (in the sense that you don’t need to create yet another account), it’s easily indexable (there are plenty of web UIs for mailing lists), it’s convenient and highly asynchronous, not to mention it’s a mature and well established open standard and decentralized protocol, with lots of open tools that fit the spirit of FLOSS in general.
Discord, however, is closed, “unindexable”, doesn’t work offline at all (with e-mail you can read and compose e-mails totally offline, it’s heavier (both in terms of computing resources and data transfer) and full of intrusive pop-ups and whatnot (and has arguably distracting money-seeking features). That’s fine and maybe desirable for certain types of communities, specially the instant aspect of it, which is a strong and harmless difference between the two, but it’s not fit for the base space for contact between developers, contributors and users.
In my opinion, of course.
-
Agreed, and in addition, I hate the web interface dependency for github and gitlab, and how many system resources they use (can’t even load gitlab on my pinephone without it crashing due to running out of memory!). At least gitlab can hypothetically have a minimal open source client. I’d much rather just communicate with developers through mailing lists. If hosting is hard, there are providers for lists.
-
I think there’s nuance to this. Of course there are asshats like MongoDB that pull the rug and enshittify; but on the other hand licenses are a tool, not an ideology. If fucking over corporations involve a more restrictive non-commercial license that isn’t open source, that’s a good thing in my eyes. It depends on the software being written and how it’s being used.
-
Fuck Discord, all my homies hate Discord; use IRC/XMPP/SMTP instead.
-
I’m not gatekeeping anything, I only care if your patches for my ports are good.
-
FSF feels like a cult, they care more about the purity of foss than its practical effects on the world; and their specific implementation of foss (copyleft). This goes back to licensing and how there’s more nuance in licensing than if it’s open source or not.
-
He missed “Drew DeVault forking maintained packages and abandoning them”
Which packages do you mean?
Maybe Redis/Redict? The development on that seems pretty dead.
As a fork of Redis there is Valkey, maintained by the Linux Foundation and licensed under BSD-3-clause.
If Valkey is the de facto successor of Redis, then maybe abandoning Redict is the right move. If he continued to put effort into Redict, people would just ask why is he wasting time with Redict when Valkey exists.
Note that I generally don’t think time put into free software is wasted, because once its put out into the commons it can be picked up and reused elsewhere - although in this particular case since Redict is licensed under LGPLv3 contributions made to it cannot be reused by Valkey which is licensed under the BSD license. One is, however, free to add their own contribution to both projects as neither requires a CLA, however both do require a Developer Certificate of Origin (DCO) https://codeberg.org/redict/redict/src/branch/main/CONTRIBUTING.md https://github.com/valkey-io/valkey/blob/unstable/CONTRIBUTING.md This is as far as I know an unusual case as generally forks use the same license allowing code to be freely exchanged between them.
