Hacker News new | ask | show | jobs
by kaba0 1043 days ago
These two are not the only two options: the nix model is clearly superior and is not as error prone as each.
2 comments

I use Nix on an ARM single-board computer to host a personal Matrix homeserver (and a bunch of bridges), and I absolutely love it. It's invaluable to have a reproducible specification of the whole system, including custom software to build, in a single place.

That being said, for day to day stuff Arch (and Nix standalone) works well enough for me, to be weary of switching my daily driver PC to Nix, out of the fear of dealing with unforeseen issues and maybe encountering less well maintained packages (there's always something broken on Nix unstable, but maybe it's not an issue for more popular stuff). So I'm sticking to Arch for non-servers for now.

nix is so clearly superior and yet nobody uses it. curious!
I think that the learning curve for the nix ecosystem is a lot steeper compared to other distros.
Well, arch is no ubuntu either.
Sometimes the better solution takes time to reach the masses.

Here we talk about a tiny percent of the tiny 3% of desktop users.

> nix is so clearly superior and yet nobody uses it. curious!

If you're going to be snide, you should at least be correct - or are the 6300 users in the main matrix chat room just hanging out for fun?

(sent from my nixos daily driver)

It's not like 6300 is anything but statistical noise compared to the number of Linux users out there.
The number of Arch Linux users is nothing but statistical noise compared to the number of Windows users; would you say that nobody uses Arch?
In that context, if, that is, the comparison involved Windows user share, then yes.

Hell, even in a Linux-only context too. I mean, an exchange like:

- We're shipping this enterprise software in packages compatible with RHEL and Ubuntu, would it be worth our while to also devote resources to specifically support Arch too?

- Nah, nobody uses Arch

While not accurate to the maximum possible precision (something like say 5% of Linux users is not the same as 0%), it would still be quite understandable...