Proposal: Moving ownCloud community discussion to GitHub Discussions

Moving the forums - fine, as long as the new locations also includes user support, not just developer discussions.

But: I think it would be more important to get off of GitHub. Microsoft is not a friend of open source.

Hi @Brad

Both points are fair and I want to address them honestly.

On user support: you’re right. Discussions skews developer-heavy by default, and “go post in the repo” is a lousy experience for someone who just wants to know why their desktop client won’t sync. If we go ahead, the org-level space needs a clearly labelled user-help category with templates, sensible tags, and people who actually answer. Otherwise we’d be moving a half-active forum to a half-active corner of GitHub, which helps nobody.

On GitHub itself: I flagged this in the original post because it genuinely bothers me too. Digital sovereignty isn’t a slogan I picked up last week. But I have to be straight about where we are. The code, CI, issues, and PRs all live on GitHub today. Moving discussion to Codeberg or a self-hosted Forgejo instance while the code stays on GitHub wouldn’t solve the concern you’re raising, it would just fragment the community in a different direction.

The bigger question (should ownCloud reduce its structural dependency on GitHub over time?) is a legitimate OSPO-level conversation. I’m watching the Codeberg/Forgejo ecosystem (I am a member myself) and the EU-side work around sovereign developer infrastructure (ZenDiS, Sovereign Tech Fund, and similar). I’m not going to promise a migration off GitHub in a forum reply, because that’s not a decision one person makes on a Sunday. I will promise the question stays on the OSPO roadmap rather than getting quietly dropped.

So: your conditions are the right conditions. User support has to be first-class, not an afterthought. And the Microsoft point deserves more than a one-line acknowledgement.

Thanks for pushing on both.

David

I think also that to get off of GitHub because GitHub belongs Microsoft and is not a friend of EU would be a great idea!

Many thanks!

Hi JClaude,

I answered most of this in my reply to Brad just above, so I won’t repeat it here. The short version: concern is legitimate, but forum-move and code-hosting-move are two different decisions and I don’t want to conflate them.

One thing worth adding on your specific EU framing: the honest answer is that no major code forge today is both fully EU-sovereign and operationally ready to host a project of ownCloud’s size and integration surface. That gap is exactly what initiatives like ZenDiS openCode are trying to close, and it’s a space the OSPO will track actively rather than from the sidelines.

Two clear +1s on this point in the thread is a signal. Noted.

David

Hi Hammerhead,

Honestly, this is the most important thing anyone has said in this thread, and I mean that.

The “where do conversations happen” question is relatively easy. The “how do non-technical users get treated when they show up” question is much harder, and you’ve put your finger on exactly why. RTFM culture is real. GitHub’s default tone skews developer-to-developer, not volunteer-to-volunteer-who-just-wants-their-desktop-client-to-sync again. If we moved everyone over and the result was novice users being told to read a man page, we’d have done harm, not progress.

So here’s where I land: if we can’t commit to a user-help space that’s genuinely welcoming (clearly separate category, plain-language templates, named people who actually answer, and moderation that doesn’t tolerate “just read the docs” as a reply), then the move isn’t worth making. I want that to be a condition, not a footnote.

One gentle push-back though. The behavior you’re describing isn’t really about Discourse versus GitHub. It’s about culture. A hostile culture on Discourse hurts novices just as much as a hostile culture on GitHub. Platform choice makes it marginally easier or harder. The actual work is on the community, and that’s something I’d want the OSPO to own explicitly, wherever this lands.

Your comment isn’t “just one man’s opinion.” It’s exactly the kind of feedback this thread needs. Thank you for taking the time to write it out.

David

Hi Roke,

Three good technical points, so let me engage with them specifically rather than waving them off.

On labelled issues as a Discourse substitute: I’d push back here. Issues and discussions are different primitives on purpose. An issue carries the implicit contract “someone will triage this, close it, or ship code against it,” and overloading that channel with support questions and open-ended community chat breaks the signal maintainers rely on to actually release software. Projects that go this route tend to end up with either a label-pollution problem or maintainers who quietly tune out their own issue tracker. Neither ends well.

On threading: you’re right. Discussions gives you flat plus one level of nesting, and that is genuinely less expressive than Discourse’s linear-with-references model. I won’t pretend otherwise. Whether it matters in practice depends on how the forum actually gets used, and looking at current activity here most threads are flat anyway. For the conversations that do branch, the loss is real.

On free Discourse hosting: yes, the FOSS hosting program is a real option, and raising it is fair. But hosting cost wasn’t actually the argument for moving. The case in the proposal is about gravity: contributors and most technical conversation already live where the code lives, and a parallel forum creates a second place people have to remember to check. Free hosting doesn’t address that fragmentation, it just changes who pays the bill. It’s a clean answer to a question we weren’t really asking.

Vote registered. The threading point in particular is one I don’t have a clean counter to, and I’ll sit with it rather than dismiss it.

Thanks for the pushback.
David

Thanks for weighing in on the pro-move side.

The reversibility point especially is worth pulling out, because I don’t think I emphasized it clearly enough in the proposal. Platform decisions aren’t one-way doors. Discussions content is exportable, Matrix already runs in parallel as a chat channel, and if the GitHub situation shifts materially the content can move. That changes the risk calculus: we’re not locking ourselves in, we’re making an operational choice that can be revisited.

On the “not a big deal” framing I want to be slightly careful though. For some users in this thread it genuinely is a big deal, and I don’t want the pro-move voices to sound dismissive of that. It’s possible to think both “this move makes sense today” and “the underlying sovereignty question deserves serious OSPO attention tomorrow.” Those aren’t in conflict, and I’d rather hold both at once than pretend the second one away.

David

Hi Terence,

This is the most coherent anti-GitHub argument in the thread, so let me engage with it properly rather than dodge.

The sanctions point is the one that deserves to be raised more often than it is. GitHub has blocked users under US sanctions regimes, and those cases weren’t hypothetical. For a project whose mission includes digital sovereignty, having contributors potentially locked out by decisions made in Washington is a real structural risk, not a theoretical one. I won’t pretend otherwise.

The mission alignment framing also lands. “ownCloud exists to reduce third-party dependencies, so consolidating on Microsoft is backwards” is a genuinely uncomfortable argument to sit with, and I’d rather sit with it honestly than wave it away.

Where I’d push back gently is on scope. What you’re proposing (consolidate somewhere, but not GitHub) is the right shape of argument, but in practical terms it’s not a forum decision, it’s an entire-project migration: CI, releases, repositories across the org, PR automation, the bot ecosystem, third-party integrations, contributor notification habits, documentation links across a decade of content. That’s real multi-year work. And it has to be planned around the actual maturity of the destinations. Codeberg is promising and I watch it closely, but operational readiness for a project of ownCloud’s scale is still being proven. GitLab SaaS is US-owned too, which is a lateral move on the sovereignty axis. Self-hosted Forgejo or GitLab solves ownership but swaps it for operational burden the project would have to carry.

So where does that leave me? Honestly: your analysis of the destination question is stronger than my framing. If ownCloud is serious about sovereignty as mission, exit from GitHub needs to be a tracked, funded workstream with a realistic timeline, not a rhetorical posture. That’s an OSPO-level commitment worth making: the question gets a roadmap, not a shrug. What I won’t do is collapse the forum decision into that roadmap prematurely, because a half-executed migration of discussions to a platform that isn’t ready to host the project yet would be worse than either option on the table.

Paid Codeberg as a middle path is worth looking at seriously. I owe that one more than a dismissal. Thanks for the sharpness. This is the kind of pushback that actually changes how I’m thinking about it.

David

So, I haven’t really used OwnCloud in a while, not since I thought OwnCloud server was being replaced by Infinite Scale and how that would work with the apps and migrating to it wasn’t clear to me. TBH, that still isn’t clear to me, but I haven’t really looked into it in depth…

I’m only replying because the first pull request someone accepted from me was to OwnCloud, so I still want to see it succeed. And I’m very very leery of Microsoft.

Microsoft’s push to gain access to every piece of data it can, so it can either advertise based on it, or train the LLM’s on it, is not something that makes me trust anything they have their hands in. GitHub is owned by them. So, if anything, I’d say moving away from GitHub is the better idea.

That said, I get why you’d want to not have to maintain Discourse. I tried running an instance a while back and it wasn’t my favorite thing to deal with.

I also understand having to have a presence of on GitHub just because so many developers are there already. I think we’re collectively crazy for that, but I get it.

Maybe just throw up an instance of GitLab’s omnibus Docker image and have a specific repo for support requests? I’ve been running one at work, and one at home, for years now. It is a bit heavy and can be a bit slow, but if you don’t try and skip to many versions, upgrades are smooth.

Anyway, just a drop in thought. Maybe I’ll see if I can figure out Infinite Scale….

Dear David, dear all

Hi everyone

I fully understand the proposal and all other things equal, it seems a sensible move. It it allows Owncloud to concentrate its resources where they are most effective - on OwnCloud! - helping me and millions of other users keeping their private data private and save. It is more than any other software company has ever done good to me.

Thanks for informing and transparency though, please keep up the good work.

Hi Natureshadow,

The strength of feeling is noted, and I want to engage with the substance rather than either capitulate or brush it off.

You draw a distinction I think is fair: that staying in big tech walled gardens is already questionable, and that actively moving community there is a worse version of the same problem. That’s a more defensible position than a blanket “GitHub bad” argument, and it deserves a real answer.

I don’t think moving forum conversation to a place where code, issues, and PRs already sit counts as “moving community into GitHub” in the way you’re framing it. The community is already there, operationally. The forum is the outlier, not the code. Whether consolidating those signals onto one dependency makes the dependency materially worse, or simply makes the existing dependency more visible, is a question I don’t have a clean answer to. Reasonable people can land on either side.

Where I’d push back harder is on “inexplicable.” The proposal has stated reasoning, acknowledged trade-offs (including explicit acknowledgement of the Microsoft-ownership tension in the original post), and a fourteen-day open consultation before anything happens. You can think the reasoning is wrong or the trade-offs weighed wrongly, and those are fair positions worth arguing. But “inexplicable” collapses the distinction between disagreement and bad faith, and I don’t think that framing serves the conversation. Several people in this thread have disagreed with the proposal in ways that are sharp but recognizable as engagement, and I’ve tried to engage back on the same terms.

The sovereignty concern isn’t going away and isn’t being minimized. I’ve said elsewhere in the thread that exit from GitHub as a long-term OSPO workstream deserves a roadmap, not a shrug, and I stand by that. But the binary framing (“do this and you lose all remaining credibility”) doesn’t leave room for the discussion that needs to happen: about timelines, destinations, and how to do this well rather than whether to consider it at all.

Your vote against is registered.
David