[svin_bee]

Has the situation between you guys and the Nix community cooled down yet?

I have honestly avoided NixOS for quite a while because the situation seemed "unstable". I don't particularly care for the complaints brought about by the open letter, I do however worry if a distribution/community is imploding.

[grhmc]

There is a lot of turbulance in the project, but I think a lot of that is a natural result of twenty years of history meeting standout commercial success. Our approach here is to continue our pattern of improving Nix and sending patches upstream. We'll also continue improving the ecosystem around Nix, most of that being open source, and only a couple specific projects being proprietary.

Regarding the stability of the project: The project and distribution is incredibly vibrant, and not going anywhere. There are tens of thousands of organizations that depend on it every day, and projects that are funded on the time scale of decades. It's here to stay.

[dandellion]

As someone who just went back to Arch after using NixOS as my daily driver for more than six months, what's in dire need of improvement is the documentation, the rest of the ecosystem was all right.

[grhmc]

Agreed. Thank you for saying so. We're working on this in two ways:

1. Reducing the number of moving parts, so there are fewer interfaces that need to be documented and understood ... because they don't exist.

2. Resources like https://zero-to-nix.com/, and https://docs.determinate.systems/.

This is a big lift, though, so anything we do here in the near future will definitely not be enough.

[dandellion]

Those seem like good steps in the right direction. Other things I'd like to see when I try Nix again would be more consensus of what are good conventions to follow for packages and configurations.

Zero-to-Nix looks better than the documentation I used to get started, seems a great resource. But even with the worse documentation I found, I was able to overcome the pretty steep initial learning curve. What did it for me was the lack of documentation for many packages and having to dive into the code every time I needed to configure anything remotely advanced. Whereas in Arch there's always a couple examples that get me 95% of where I want to go, making it very easy to figure out the rest.

But thanks for the effort, I'll be looking forward to giving Nix another go in one or two years to see the progress and maybe switch again, as I really liked the whole concept.

[marliechiller]

I'm still driving nixos, but for me, the stacktraces are the hard part when discerning where and why things have gone wrong

[grhmc]

Ugh. I feel you. I think one thing that makes this hard is how big and intertwined Nixpkgs and NixOS is. All the modules are loaded all the time, and it can make the displayed errors propagate out to somewhere intractably far away from where the error actually is. There's a ton of work still being done in Nix to make that better, and I'm hopeful for an easier future :').

[deng]

> There are tens of thousands of organizations that depend on it every day, and projects that are funded on the time scale of decades. It's here to stay.

I like Nix, but what turns me off most is this kind of hyperbole from its evangelists. "tens of thousands of organizations" depend on Nix every day? How do you even come up with that kind of number? And as someone who has actually worked on projects that are funded for decades: the first rule is that things are changing, so stay flexible, chose boring technology, and always have a plan for migration. Binding yourself to a niche framework that is also strongly opinionated with an arcane syntax is pretty much the worst you could do.

As you say, Nix is already over 20 years old, so yes, it's probably not going anywhere, but that goes both ways: it will probably not vanish in the near future, but it'll probably also not become really popular. And that's fine.

[callahad]

I think it's worth jumping in. Fifteen years ago, it wasn't clear whether Git, Mercurial, or Bazaar would win, but it was evident that something fundamental had forever changed in revision control systems.

Nix feels similar, but for build systems, package management, and configuration management all at once.

[dietr1ch]

Right, I don't know if nix will manage to evolve, or a new implementation will come up, but the box with the wonders of nix has been already open and there's no way people will want to go back to mutable and non-deterministic systems.