NixOS is not dying, please don't spread fear actively

In the big discussion about organizational structure that started again a few days ago, we have initially reached consensus on some aspects. However, some unnecessary and sensational news has been spread inside and outside the community, which, imo, is destroying the reputation NixOS has built up in this area through concerted efforts, deter new users from joining NixOS, and spread fear to some existing users who are not yet fully aware of what’s happening.

NixOS is not dying, NixOS is not collapsing, NixOS is not becoming bad altogether because of some decisions with lack of consensus being made. Why? Hear me out.

The main part of NixOS, the GitHub - NixOS/nixpkgs: Nix Packages collection & NixOS, the product of more than two decades of hard work by thousands of maintainers, should not be seriously affected by the status quo. I’ve explained in another topic yesterday, but let’s just be clear:

  • Nixpkgs/NixOS is driven by Nix, that’s right. but we are still actively maintaining compatibility with Nix 2.3 and have not made any new features necessary to use, whether it is flake, or nix command, or IFD (Input From Derivation). Thanks to the efforts by The Virus Lounge, Tvix is ​​currently basically compatible with Nix 2.3. Therefore, Nix is ​​not a requirement of Nixpkgs. No one can hinder the development of Nixpkgs as a Nix developer.

  • Nixpkgs’ public resources, such as hydra and cache, are donated by volunteers and IT companies. This donation is not due to the leadership of the NixOS foundation or anything else, but because of everyone’s recognition of the robustness of NixOS and the urgent need in NixOS development in this regard.

  • The “control” of Nixpkgs is not in the hands of the foundation, moderation team, or someone, if any, who implicitly control the organization, but in a team of committers composed of more than 200 people. Why do I put quotation marks for “control” here? Because we are not actually exercising this power, but acting as a service provider to solve problems for everyone. This is explained in more detail in the reply cited above, but I missed one thing in this reply. The conditions for applying for a committer are jointly controlled by the inspection and comments of everyone (all those who can comment in that GitHub repo), rather than by a few people. (If you are experienced, have enough time to waste, and want to help others, please don’t hesitate to apply!!)

  • Nixpkgs is a success of community participation, not a success carefully designed by a few people. Thousands of maintainers make up a majority of the active community, and Nixpkgs is the beginning of making everyone feel included. It’s not that there aren’t arguments or ill-considered decisions, but it’s very different from other parts of the main problem we face.

  • Regarding the issue of “representation” or “endorsement”, if we think who should not represent us, we should actively solve this problem, instead of directing the problem to the remaining majority of “represented” people. It’s no use.

I once again explain to people who are not fully aware of this and are frightened by this incident, that these arguments are not terrible. On the contrary, we are actively removing stumbling blocks on the development path of NixOS. We are forming a new, more complete consensus based on the old consensus to covers more users, and have more respect for community opinions, thereby forming a more closely united community. I believe we will succeed.

I hope our strategy is not to do bad things to force everyone down, in order to ensure that everyone will “face this problem head-on” together, in case some people really think so, or their behavior is driven by this goal. This is unlikely to succeed and will only lead us to a worse outcome. I hope that these behavior, whether it is based on a herd mentality, a speculative mentality, or a sabotage mentality, please stop. We don’t want to see “NixOS is dying, here’s what you want to know” or “Why you should leave NixOS” again.

At the same time, I need to point out for @wegank again that NixOS 24.05 is near, and Zero Hydra Failure Compaign is requiring everyone’s participation. I believe we can continue to fight side by side, both in community affairs and project maintenance.

For people who are already exhausted, I suggest you to run nix run nixpkgs#blahaj -- --shark, and get the blahaj I’ve prepared for you <3

137 Likes

Thank you. I’m super optimistic about the future of NixOS. Don’t get dragged into negativity. We’re more popular than ever, this comes with large challenges, but we will solve them all. I’m hopeful and 1000% committed to make us succeed.

51 Likes

Zero Hydra Failures nominally starts May 8th. it’s one of the more accessible ways to contribute to the release cycle: we’ll have a list of package build failures, choose one of them (or two, or ten), and get it building again. i plan to participate in this one more than the previous ones specifically because in the context of everything else, this seems like an opportunity to make things less stressful for the people who bring me to this project in the first place.

(May 8th - May 30th: put it on your calendar or watch for the announcement!)

19 Likes

Nit: this is not 100% true. Some of it is just paid by the foundation, for example.

But I wouldn’t focus on details, and I do want to encourage the positive thinking. Sure, not everyone around NixOS.org has the same opinions and values, especially around “political topics”. Unnecessarily heated discussions got sparked, etc. Of course those aspects aren’t pleasant, but that certainly won’t deter me from usual technical contributions.

22 Likes

The Foundation only currently manages donations and does not generate any income of its own. Yes, we will also encounter a problem if the controllers of these services do things that are detrimental to Nixpkgs. But what I want to express is that we have reached this point entirely on our own, and we are confident that we will continue to receive support. It is not that we are dependent on others, but we are playing a leading role.

Theoretically, given the above structure, how doable would it be to jettison the current foundation, construct a new one, and perhaps shift to something like the rust reimplementation of nix?

Compared to a fork, that would seem a lot less involved as extreme measures go?

1 Like

It’s doable as a last resort, but it still break a lot of things and consume existing resources. I still hope we can solve the problem through discussion and negotiation first, and that doesn’t seem impossible.

8 Likes

Sounds like pretty extreme fork of the community to me.

7 Likes

Absolutely. An extreme measure should by its very nature not be preferable.

But it is an interesting datapoint vs. “a fork would mean new infrastructure, so it could never work, exiting the project altogether is the only viable extreme measure”.

2 Likes

Absolutely. My shorthand “fork” was intended to refer specifically to forking code resources.

1 Like

The starting point of this discussion is to demonstrate that “we play a leading role”, and it does not mean that we will fork immediately. Currently shouting “we need to fork” does not help matters, and is prone to the problems I describe in this topic.

15 Likes

If you are interested in this i suggest hitting up flokli who is working on tvix. It’s quite a serious attempt at doing this.

I for one think having two parallel implementations of nix is great. Without abandoning the one or the other. It solidifies the existence of nixpkgs and makes our community stronger. To stay on topic :slight_smile:

19 Likes

Agreed. I’m still here because some things I care enough about to fight for (and I have the energy to do so, enabled by personally having it easy in the community).

But in attempting to problem solve a primary asset is clarity of information. With the understandably increasingly passionate discussions, such is hard to come by. Thus, when you provide it in such high grade in this thread, I seek to make the most of that opportunity for exploration.

2 Likes

100%

Irregardless of how this all pans out, solidifying the standard through parallel implementation is IMO paramount for long term success.

1 Like

This is both awesome and well timed IMO <3

8 Likes

Thanks for writing this! I just want to quickly let everybody know that I won’t go anywhere, and that I am very optimistic about the future. Yes there are problems, but we can fix them, and I’m talking with various people from the community about how to do that. :heart:


Thanks to the efforts by Tweag, Tvix is ​​currently basically compatible with Nix 2.3.

Brief note: Tweag doesn’t currently do any work for Tvix as far as I know! I guess you’re thinking of Re-implementing the Nix protocol in Rust, which is unrelated to Tvix :wink:

47 Likes

Who is “spreading fear actively”? Major contributors are leaving the project. People are fed up. Pretending it isn’t happening and kicking the can down the road won’t improve anything.

9 Likes

Yes, NixOS is currently taking a big reputational hit in terms of public perception. And I think that this is not only warranted, but neccessary.

It is warranted, because people who are new to this community or want to be involved should know what they are getting themselves into. That’s just informed consent.

It is necessary, because all other means of resolving the issues through internal means have failed or not been sufficient. People are now voting with their feet, and this creates the necessary external pressure to actually do and improve things. The first open letter shouldn’t be ignored, but it could be ignored. On the other hand, this can’t be ignored.

26 Likes

You mention that major contributors are leaving and that people are fed up. What do you think that that means for the future of the project? I ask because it sounds like you disagree with @Aleksanaa’s post, and I want to understand better why you disagree with it.

1 Like

About 20~30 contributors, some used to manage some of the important parts of Nixpkgs, are leaving the project. This is fact.

  • Maintainers are leaving
  • Major maintainer is leaving

These are not. They are indeed true in a sense, but to others who have not witnessed the entire event, it expresses a meaning that is clearly not true.

As a side note, we currently have 2900 maintainers registered as Organization members, the actual number registered in maintainer-list.nix may be higher or lower than this. According to Repology, we have 3284. At the same time, depending on the way and purpose of using Nixpkgs, we probably have many major aspects. We have 78 topic labels. We do not currently have an accurate measurement standard for people who have made “major” contributions to Nixpkgs and are still actively contributing, but the number should be no less than 30.

I feel sorry for some of the current maintainers making the move to leave Nixpkgs. I don’t want to measure their actions with any bad intentions, and I understand the actions they took.

But I do not accept these misleading titles. They describe the current situation in an unnecessary much more serious way. No matter what the purpose of this behavior is, it is causing panic that is much more serious than the matter itself.

21 Likes