Of course, everything we are talking about in regards to Nix is speculation right now. It remains to be seen if the right kind of governance can be established. And it remains to be seen if the other problems of the Nix ecosystem can be addressed. For now I think it’s best for us to continue with Aux while trying to guide Nix in the same direction.
well it’s great both ways, because either Aux becomes a really good and healthy fork of Nix’s ecosystem or we get to still all get together and make a technology we all care about just… work better. Aux has value either way. maybe it becomes a more organized “nix-community” sort of thing, i don’t know. but i’m really looking forward to the future of Nix overall once again.
Hello Aux
I figured I’d pop in to let you know I appreciate this post.
I’m on the team that maintains the package manager on the Nix side. It’s probably early to talk about changing things about the package manager, but whenever you’re ready, I’d love to talk and keep in touch. I believe that could be quite helpful to get you started, and good communication always leads to a better outcome.
Thank you for reaching out @roberth! As things progress we could definitely use all the help we can get, especially the package manager. Looking forward to chatting in the future!
I believe that the best course of action right now is to do… both
I firmly believe this as well.
And to extend that idea, Nix and Aux could be symbiotic. There is an old and long thread in the nix discourse on unmaintainablity with discussions about things like merge trains and other strategies. If Aux intentionally tries alternatives, and happens to find one that works a lot better, maybe it will get ported back to Nix. And potentially vise versa.
I have long worried about the unmaintainability of the monorepo, so I was actually pretty excited to try and reopen the discussion in the Aux discourse since there will be a smaller team handling maintainance. A community with fewer members can often coordinate easier and move faster, even if it can’t perform heavyweight maneuvers.
Excellent point, it would be interesting to see how forking with a smaller team might actually have some benefits compared to the Nix powerhouse. Maybe it could lead to quicker innovations with things that Nix is otherwise too inflexible to coordinate on given its scale and governance issues.(fingers crossed there will be significant progress in reforming it soon, of course)
Theres also lots of different directions. Nixpkgs the repo has a lot, a LOT, of legacy. This project could be one that falls back on nix (for top-level.nix) but otherwise tries to be a streamlined flakes-first redesign of nixpkgs.
- Formatting required for PR’s
- Flat package structure rather than a hand picked heirarchy (small team can’t hand pick, must just be systematic process)
- Package requirements (must have pname and name and version not “maybe one, maybe two” of those things)
- etc
So like a “beta” version of Nix, where we move fast and break things just to see what sticks? I think that would actually be really cool.
I was really shocked to wake up and over night see that the NixOS foundation put out a notice to the users. And I think it is a step in the right direction for them but not quite reaching the goals that aux was striving for. I feel like continuing with aux is a good idea for now. And maybe one day like io and node we can rejoin if they can meet our goals.
this post has just become the “should minorities be protected” post again, with people arguing about sea lioning, and 0 moderator action taken at all
I’m not really convinced by the board’s response, but I guess I will wait and see what the constitutional assembly achieves… If it doesn’t get derailed and trolled like the response’s thread, that is–
So for now, nothing changes for me, I’m still fully committed to helping Aux succeed.
Hello there
Thanks for this post. I did sign https://save-nix-together.org/, therefore I will support Aux. With this post, I understand that me staying in Nix community for now, keep working on nixpkgs, and keep advocating Nix to the outside world are actions that support Aux.
Please don’t hesitate to correct me by any mean if this is not the case, or when this will no longer be the case.
I strongly agree with this statement (and the whole post).
I’m happy to see that the Foundation has taken a step in the proper direction and that there’s willingness to continue improve.
While it’s just a start and “”“nothing”“” is yet done it’s promising.
For the time being I think it hurts noone to continue working on this alongside nix.
it would be cool if Aux was sort of like another OS to NixOS as Pop!_OS is to ubuntu.
Getting a bit off-topic but still wanted to say
I think SnowflakeOS is closer to Pop!_OS in that analogy though.
Aux seems closer to systemd-less forks but instead of big technical change it’s a big social one.
I agree, since Aux should eventually not depend on anything Nix. As for SnowflakeOS, right now I plan to keep developing It with Nix until Aux matures, then potentially swap over (especially if I can improve Aux package metadata).
Aux has value beyond just leadership changes. The discussion on threads in this forum suggesting changes to things that simply stagnated for years in the original Nix ecosystem, which everyone came to accept as the “has always been” situation, was the main reason I was excited about Aux in the first place.
It is unfortunate that both resources and man-hours are scarce, and these changes must be implemented gradually to ensure the functionality (heh) of everyone’s system and packages.
Aux provides both the ability and the reason to work fast, without devoting so much time to backwards compatibility.
We have already identified and can improve the nearly-vertical-learning-curve experienced by new Nix users.
While this is a bit contrary to my previous statement, in the absolute worse scenario, Aux will serve as a hypothetical, a what-if Nix did …?
This should be of interest even to members of the original Nix ecosystem.
Considering the nature of the project, should its need eventually disappear, we can learn enough (both successes and mistakes) from it for it to be worth doing.
Aux provides both the ability and the reason to work fast, without devoting so much time to backwards compatibility.
I agree with everything you say but this bit. We need to dedicate some time to backwards compatibility at least for know whilst we are in this transition phase since we want to help users swap with ease. With limited people we can only dedicate so much effort towards making certain changes.
I am supportive of Aux, but at present I am investing my time into trying to help Nix out of the present difficutly. Should I request to become a member of the org now or wait until I have more time / the Nix situation is settled?
You can do whatever you like. We have a post for requesting access GitHub Organization Membership, though I would prefer those numbers to be people that would actively contribute such that we have a rough grasp on how much people power we have.