Could the name become a problem as aux is a reserved file name on Windows?
How much of an impact would this have on WSL? I guess this is a non-issue if Windows isn’t a concern to the new project.
Nice to see we’re getting a bit of a headstart on spinning up a fork. The overall idea sounds like a good foundation to me - if we end up having to fork, this seems like one I’d be willing to back, at least based on the (admittedly basic) information presented here.
Personally, I’d like to see change happen here, if possible, over a fork - given we’re still awaiting a formal statement from the Foundation, I would probably want to hear that first before fully committing to any fork. Forks tend to be quite divisive last resorts, with a lot of effort for potentially little to no payoff, so if we can fix things in the community we have now, I’d consider that a more optimal route than forking. Whether that’s possible is yet to be seen (which is why I’m looking forward to hearing the Foundation’s thoughts on the matters), but I guess I’m a bit of an optimist. I think there’s still a way forward to mending the broken processes and structures here - I just hope we can commit to putting the effort in.
21 Likes
While the huge technical efforts of a hard fork and the resulting divergence make them somewhat counterproductive, I think that alternative Nix communities could be a huge win. There are some risks (potential for additional drama between communities, fragmentation, etc.) but I think that the people behind aux.computer should seriously consider that it might be a good idea to at least try to foster a community still even if the desire to fork fades. Some common ground for contributors to the same project would always be needed for collaboration, but even in the absence of a hard fork, alternate communities could alleviate some of the stress caused by differing social norms among people of different perspectives.
To me, this is the logical answer to the existence of a “determinate systems” community. Whether or not forks and separate infrastructure needs to exist is an important question for which the answer may not be fully realized, but yet I also believe it is somewhat beside the point in the grand scheme of things.
I understand the fear people have with forks, but I think we (as a group of Nix enthusiasts and fellow humans) should not fear them. Hard forks should not be done lightly, but that they can be done is a huge part of what makes free and open source software so powerful.
I care more about Nix itself than any of these conflicts, so I ultimately plan on trying to follow whatever seems to be the future at any time, so as long as I am allowed to do so.
5 Likes
@jakehamilton Is there some way to get notified of any updates to this project? It would be nice if this page linked to a feed for those to follow such as Atom, RSS, ActivityPub, and/or XMPP PubSub.
16 Likes
While I agree with forking nix and (probably) nix-darwin for governance reasons, I’m apprehensive about the proposed nixpkgs changes because I am doubtful Darwin support will survive.
I don’t think my concern is unfounded considering there was an RFC to demote Darwin to tier 3, which failed, and another call last year to revisit demoting it again.
If a core SIG can decide they aren’t interested in Darwin support, then that’s it? If they do decide to support it, but other SIGs don’t want to be bothered, then what? That’s why I asked about how project governance will work.
6 Likes
I must admit that I agree your fear is definitely founded in my opinion. Regarding that specific issue though, I can simply only imagine it’s too early to say (and honestly, I sort of hope this thread doesn’t derail into a long recap of the pros and cons of Darwin support.) If it’s important enough, Darwin support will surely help decide which fork is most successful, I can only imagine.
(Also, my intent with replying was not to shut down any discussion or suggest there’s nothing of concern, I just wanted to express that I think these developments could wind up being very positive, in response to a lot of the fears.)
1 Like
I’d think it’d be the other way around. If a SIG of significant size can form for Darwin support, then there will be Darwin support. Lets be honest, any fork will start out with significantly fewer resources than nixpkgs, even if everyone who said they’d contribute will actually stick around for longer. A lot of work will be spread very thin for a long time, and Darwin would be a LOT of work added that is not easy to distribute due to its proprietary nature. I cannot just spin up a VM and test changes locally or run a CI builder for it.
So I agree, it’s probably not looking to great for Darwin either way, but that’s not due to a lack of governance or motivation. If Darwin support does not happen or gets dropped a tier on nixpkgs, it’s because not enough of people are available to do the work.
1 Like
I mentioned it to contextualize my worries, but I agree. That conversation should happen elsewhere.
Platform concerns (whether Darwin or Musl or FreeBSD or Windows) tend to be cross-cutting. A platform SIG needs to be able to set standards the other groups can be reasonably expected to follow, or the others at least need to defer to them when it comes to platform needs or accepting fixes for platform support.
Platform concerns (whether Darwin or Musl or FreeBSD or Windows) tend to be cross-cutting.
I agree and I am a (or rather: was) trying to improve Musl support in nixpkgs, but it kept being a Sisyphean labor and in the end left me quite bitter to the point I stopped contributing much at all. So I can sympathise, but specifically because these cross cutting efforts, we need to be realistic about what can be practically maintained and consciously opt in to the work with a clear idea of who’s going to do it
2 Likes
This is an intriguing development…
To be clear, a Darwin SIG would be established if there is interest. If members of that group no longer find it interesting, they can leave but elections will resume and people who are interested may join. The only way that Darwin support dies is if there is absolutely nobody willing to support it.
7 Likes
You can check back later today, I should have a forum up and running. I want to make sure it’s got things organized and ready first. There are a lot of people who have expressed interested in joining 
I will post in this thread when that happens.
18 Likes
Thank you for putting the foot down and coming out with a solid declaration of values and goals. This is something that Nix community needed for a very long time.
It only just occured to me that the second-to-last paragraph of the open letter is very ambiguous and has multiple readings. As a signee of the open letter, I read it as “if no satisfying action is taken, on May 1st we make the letter more well-known and commit to supporting a fork”. As such, I’ll commit to supporting a fork if on May 1st, nothing satisfying comes from the upstream side.
If I may, I have a couple of criticisms regarding the goals and the roadmap. Some of the goals are a lot more constrained in tech or people, rather than in governance. The CLI stabilization effort, for example, is more constrained on people’s capacity. I don’t think it is quite time to discuss the specifics, I just wanted to point out that not goals are inherently reachable just from better community structure alone.
Regarding the roadmap, I think you’re vastly underestimating the role infrastructure currently plays. I think we’ll need to set up our infrastructure much earlier than even electing governance - binary cache is of utmost importance, and not having it would grind any progress in terms of actually submitting changes down to a halt. We will also need to establish a legal entity, and apply for financing and credits. This is so important that I think it has to be largely resolved in the first month, otherwise the fork will have a very hard time being successful.
Additionally, communication is very important. I think it would we valuable to cooperate or communicate with folks from SpectrumOS, who have maintained something similar to a fork of Nixpkgs for a while. I think there’s a lot of valuable things to be learned from their experience, and capitalizing on that existing experience would be great. And I think a number of recent Nix initiatives looked quite similar to SIG+WG model - there’s also much to learn from that. And, well, you really need some platform for discussion - having a Matrix group would probably help the initiative get more attention.
This is my criticism, and I believe it to be founded in reality, but don’t let it discourage you. I see a lot of value in this effort even if the current governance crisis will be resolved without a fork. In that case, you’ll still create a community that emphasizes improving governance and fostering good community. That will be valuable experience, and even if the fork ceases to exist, I am optimistic about being able to use that experience to improve Nix community as part of an SIG.
11 Likes
I have created a forum here for the project:
11 Likes
I’m not sure why it’s being assumed here that the authors of the open letter are the same people involved in this fork.
5 Likes
Jake signed the letter, made this post, posted to his mastodon.
I am not an author of that letter, only a signatory.
1 Like
I am not sure that changes very much. If we’re being pedantic, the letter doesn’t even say anyone will fork, only that they’ll support any fork; I assume this falls under “any fork.”
1 Like
It’s clear there is some alignment on values and goals, and that this fork hopes to welcome signatories and other supporters of the most recent open letter if they’re interested.
I think that’s as far as any connectiom goes.
3 Likes