Fair, and noted. “It just works better as a forum” is actually a more important data point than it might sound, because Discourse was built specifically for this shape of conversation and Discussions wasn’t. Preference from someone who uses both is worth more than abstract feature comparison.
Vote registered.
David
Hi jerrac,
Glad you chimed in, and the “first PR accepted was to ownCloud” line genuinely matters to me. That’s a relationship the project owes something to.
Let me take your three threads one at a time since they’re actually different.
On the oC10 to oCIS confusion: that’s not on you. The migration story has been muddier than it should have been for a long time, and contributors like yourself drifting away over it is a community-communication failure worth naming out loud rather than glossing over. Short version: oC10 and oCIS are both still around, oCIS is the forward path, the app ecosystem looks different by design (smaller surface, different integration model), and a readable migration framework does exist but hasn’t been surfaced anywhere near well enough. Happy to point you at something concrete if you want to take another look, or drop it in a separate thread so it doesn’t derail this one.
On the Microsoft-AI angle: you’re raising something the earlier comments didn’t, and it deserves to be flagged separately. The sovereignty concerns in this thread so far have mostly been about jurisdiction and sanctions, which are structural. What you’re pointing at is behavioral: LLM training on public code and issues without meaningful contributor consent, and the general data-extraction posture. That’s a different axis and it’s getting worse over time, not better. I don’t have a tidy answer for it and I won’t pretend otherwise.
On self-hosted GitLab omnibus: you’re speaking from operational experience, so I’m not going to pretend upgrades are harder than they are. But for a project the size of ownCloud, running an omnibus instance just for support discussion while code, CI, and releases stay on GitHub would reproduce exactly the fragmentation this proposal is trying to reduce, just with extra infrastructure to feed. Moving all of that is a different decision, and it’s the one Terence and others are pushing toward. Not necessarily a bad destination to consider, but multi-year work rather than a forum move.
If you do poke at Infinite Scale again, let me know how it lands. Honest feedback from someone who stepped away and came back is worth more than ten smooth-path case studies.
David
Thanks Chris, that means a lot.
The “concentrate resources where they’re most effective” framing is worth sitting with for a second. It’s easy to treat infrastructure choices as neutral plumbing, but every hour spent maintaining parallel community surfaces is an hour not spent on the thing that actually matters to users: the software working, staying private, and staying yours. That’s the axis I care about most, and it’s the one I sometimes worry gets lost when platform debates get heated.
On transparency: that’s the part of this whole process I’d defend hardest regardless of where it lands. A proposal with a fourteen-day open window, trade-offs named up front, and every response taken seriously is how these decisions should work in open source. Whether the conclusion ends up being “move,” “stay,” or “something else entirely,” the process is the part that has to hold up. Glad it reads that way from your side.
David
Ah, true.
I didn’t explain what I mean fully: I’d move all development to GitLab. But that leaves the ease of access GitHub provides behind, so…
GitLab does have some ways to mirror data between GitHub and itself. But I can’t recall how it all works. I tried using it, but our work instance is not exposed, and some of the functionality might have been behind a paywall. I just know it didn’t work as well as I wanted when I tried. That was a few years ago though.
Everything is a trade off, and it is hard to figure out what the best trade is. I just look at how much power big tech has, enough to buck governments if they want to, and I really don’t like it. I mean Microsoft has more users than most countries have population, so…
For the record, I’m currently moving my mother to Linux. Dad actually moved years ago. So I am putting in at least some effort to get away from all the big players…
(I did poke at Infinite Scale. Didn’t find an easy to use way to throw up a test instance in the docs I found. The docker compose info I found was very much aimed at production and not a quick test, and I haven’t spent much time trying to adapt it. Been working on improving my self-hosted email server instance…)
Hi jerrac,
Thanks for clarifying. “Move all development” is a different argument from “run a support forum elsewhere,” and the honest version is: yes, that’s the conversation that actually needs to happen as its own workstream. I said something similar to Terence further up the thread. It’s not a forum decision, it’s a multi-year project-infrastructure decision, and it deserves a dedicated thread with real numbers rather than getting decided sideways in this one.
On GitLab mirroring: your memory is roughly right. Pull mirroring (GitHub to GitLab) is available on the free tier. Push mirroring (the direction you’d actually want for a real exit) is a paid feature on GitLab.com, or free if you self-host. Mirroring also only covers git refs; issues, PRs, Discussions, Actions workflows, and the bot ecosystem don’t come along, and that’s usually where the actual migration pain lives. So it’s a starting point, not a solution.
On the big-tech-versus-governments point: I won’t argue it. It’s true, it’s getting more true, and it’s one of the reasons digital sovereignty stopped being an abstract talking point over the last couple of years. Moving your mother to Linux is real work in the same direction. Respect.
On Infinite Scale being hard to spin up quickly: that’s a fair hit and I’m writing it down. The docs landscape is weighted toward production deployments because that’s where paying customers live, but a “laptop in fifteen minutes” path should exist and be front-and-centre. If you want the quickest route right now: the single-binary ocis init && ocis server flow gets you running locally without Docker at all, and it’s genuinely faster than the compose examples. But the fact that I had to tell you that instead of you finding it in the first page of docs is the actual problem. Anyway, and install.sh is on our todo list.
Good luck with the mail server. That’s a different kind of masochism and I respect it.
David
I highly vote against splitting up the forum into multiple discussions in GitHub on repo level.
We only need ONE forum where users can interact, help each other, discuss about what ever - for sure not as developer only location. We devs work in issues. I really want the clear separation of user centered forum and issue tracker for developers.
But back to the forum split: one forum/discussion is enough - we cannot watch and maintain 10+ separate discussions.
Also: the main GitHub discussions is linked to the owncloud/community repo where we can maintain forum maintainer roles regardless of any other repo role setup - this is key to allow the community to step.
Apart from Discord, I find GitHub the MOST! confusing place on the internet. Move if you must but as a dyslexic person, change is never good.
I’ve made my living from Discourse for a decade now, so take this with a grain of salt. I don’t use ownCloud much these days, so that’s one more reason my opinion doesn’t matter. It does seem surprising that a Digital Sovereignty platform is discussion further tying in with Big Tech.
Are you making use of the Discourse Github plugin? It makes it easy to be able to talk PRs in Discourse, I think solving some of the problems you describe–it allows github to be for code only ad all discussion can live in Discourse, which is designed first for Discussion. It automatically links PRs at github and Discourse, making notes in both places that a PR has been mentioned at Discourse. And if you did want to leave Github one day for some other version, it’d be a pretty seamless transition. And if you one day decide that you don’t want to use Discourse’s hosting, managing it yourself somewhere else is an easy transition.
Recently Microsoft banned accounts of creators of other open source - veracrypt.
Moving ooensource to github is terrible idea.
Maybe try codeberg.org instead?
A clear NO from me.
First and foremost, GitHub Discussions is simply an inferior platform for general community communication compared to Discourse.
It is important to understand the distinction between different types of community interaction. For developers, GitHub already provides the right tools as a natural evolution of developer needs: GitHub Issues for bug reports and feature requests, and GitHub Discussions for broader code and architecture conversations among contributors. These tools grew organically out of how developers work and collaborate, and they serve that audience well. That is where GitHub Discussions belongs and where it works reasonably well.
General community discussion is an entirely different category. It includes support questions from sysadmins and self-hosters, user feedback, how-to threads, announcements, and conversations that have nothing to do with a specific repository or codebase. That is exactly what Discourse was built for, and it excels at it.
If the people proposing this change do not recognize that distinction, then the question itself is based on a false premise: you cannot replace a community forum with a developer tool and call it equivalent.
Splitting discussion “across the relevant repositories” is particularly concerning. Right now, a user with a general ownCloud question can come to one place. With GitHub Discussions spread across repos, where do they post? This creates confusion and fragments institutional knowledge across multiple disconnected threads.
Not all ownCloud users are developers. Many are sysadmins, self-hosters, and everyday users who don’t have (and shouldn’t need) a GitHub account to participate in community support. This move effectively raises the bar for participation and risks shrinking the community.
Discourse forum posts are indexed well by search engines and are highly discoverable. GitHub Discussions is significantly worse at this, meaning years of accumulated community knowledge becomes harder to find, and new users Googling their problems are less likely to land in the right place.
Moving the primary communication channel onto a Microsoft-owned platform sends a signal that the project is narrowing its focus toward developers only, alienating the broader non-technical community that makes an open source product different from just an open source codebase.
Once central.owncloud.org is archived, the vibrant and searchable history becomes a read-only artifact. Community momentum is hard to rebuild once it’s broken.
Before taking that irreversible step, it is worth asking what is actually driving this proposal. If the concern is cost or maintenance burden, there are better paths forward. Consider whether the Discourse instance can be hosted more cheaply, or whether the community itself could help maintain it. That is a far better outcome than losing a decade of community infrastructure.
Wow did not even know there was a way to so tightly integrate GitHub into Discourse through a plugin! The features sound pretty cool - not least the automatic link→permalink facility.
I’m relatively adept technically but a novice OwnCloud user and the advantages of this forum over GitHub discussions aren’t just cultural, they’re functional: the rich features (search, user & thread links, multimedia attachments) of the Discourse platform vs GitHub greatly facilitate user/developer interaction and solution discovery when creating fresh installations and troubleshooting established ones. I am involved in another tech ecosystem (Hubitat home automation) that splits out discussions and user support on a Discourse instance with the code base on GitHub and frankly it just doesn’t seem like a problem, maybe because the users/developers ratio there is higher than here OwnCloud - but wouldn’t you like to develop and accommodate a growing user/developer ratio on OwnCloud too?
I don’t carry much weight here really so take my opinion with a pinch of salt but…
As others have pointed out, the irony of a system that I use to specifically avoid MS (& Google / Dropbox) being maintained on a platform that is owned by MS and then moving a currently open source part of that ecosystem into that MS platform doesn’t sit well.
That said, I also run a business and understand that having fewer moving parts to maintain is better for productivity and increased productivity helps everyone.
The better question to be asking might be;
“Should we stop using the forums and move to github or stop using github and move it to something self hosted and open source”
I guess the issue is that from a logistical point of view it’s a lot less effort for you to move the forums than git AND users who are actively seeking assistance with their issue will go wherever you tell them they have to, particularly when that’s a free of charge resource HOWEVER getting a load of devs who already live in github to start contributing elsewehre is going to be a lot more difficult. That being the case, the “best” thing for the project as a whole might not be the best thing for the open source community.
Tough call. Don’t envy you.
Exactly this!
There’s also this, that shows the status of a PR. It’s all pretty cool, actually.
If you want ownCloud to be used only by people who know what github is and have an account, then you don’t need a forum. I think, though, you want to reach some people that don’t know what github is. I’ve supported lots of self-hosters on Discourse who don’t know what github is. I think you don’t want to leave those folks behind.
I use the Discourse app on Android to keep the forums I care about in one easy to access place. Moving the discussion to GitHub would mean that I wouldn’t read them anymore.
I think a lot of people are leaning against GitHub, and though I agree, I will try to offer a slightly different way to look at it.
For someone who thinks the right choice is for Kiteworks to migrate to GitHub, what would you advise ownCloud customers/users to do? If an ownCloud customer was in the same situation that Kiteworks is in, would you advise them to move away from self-hosted forums onto Microsoft’s GitHub? What about the decision on self-hosting forums is different from the decision on self-hosting data and all the other services ownCloud delivers?
If the question is “should I self-host X?” how is it that we arrive at “yes” when X is something ownCloud does, and arrive at “no” when X is something ownCloud does not do? Does this undermine Kiteworks’ credibility when trying to persuade self-hosting customers that they should self-host using ownCloud?
There’s credible data that GitHub has terrible uptime lately. In the last 90 days, their uptime appears to be about 87.8%. It’s so bad that Microsoft doesn’t even release aggregate uptime stats any more. To put that in context, if a web site could operate at 99.0% 11 months out of the year (333.63 days), and then shutdown and turned off for the entire month of February, it would have 91.4% uptime for the year, which is better than GitHub is doing lately. I bet you this Discourse forum has better uptime than GitHub. It’s not hard to beat such a low score. So moving to GitHub might actually reduce availability of the information here on these forums.
I hope Kiteworks sends a bit of money into the Discourse ecosystem (consulting? support?), as a side-effect of relying on this awesome bit of open source software. It’s kinda what we hope people do with ownCloud, too, right? Open source projects need to support each other. There’s a legend that Ben Franklin said about the founding of the US and fighting British tyranny:
“we must all hang together, or surely we shall all hang separately,”
I think there’s applicability to open source projects there. Everyone has to hang together, or the big tech companies pick them off one by one.