[Nix-dev] how does NixOS ensure focus?

Mathnerd314 mathnerd314.gph at gmail.com
Mon Jul 28 05:37:02 CEST 2014


On Mon, Jul 28, 2014 at 3:55 AM, hasufell <hasufell at gentoo.org> wrote:

> However, since I'v seen a lot of things gone wrong in distros (primarily
> gentoo), I am curious if you have a concept for ensuring that NixOS
> stays focussed.
>
I believe the closest thing available is
https://nixos.org/wiki/The_Many_Cooks_Method. So far, Eelco has maintained
most of the infrastructure, but in the future (according to that wiki page)
other organizations will begin maintaining their own installations as well.
Perhaps this has already started in the form of Guix; they have their own
instance of Hydra, their own Guix packages, etc., but still use Nix, the
underlying software, basically out of the box. As Ertugrul says, the
development model is very open to contributions; you just have to figure
out which repository you want to get them into. The only truly difficult
part is merging everything back together, but so far Git and the
expressiveness of Nix have been sufficient.

> Gentoo here is almost the complete opposite with >200 core developers
> and it is rapidly losing focus and consistency every year.
>
There have been a fairly steady stream of interested Gentoo users and
developers (yourself included, presumably), but I believe most of them have
turned away. They have said (probably correctly) that the project is still
a bit too experimental / unstable for them to switch. That will change in a
few years if Nix keeps chugging along, so I am not too worried about it,
but it is something to keep in mind. Eventually, they will have to be
recruited somehow into testing all of the unstable packages.

>
> I almost think that a centralized distribution model is doomed to fail.
> That does not only include the workflow with review platforms, DVCS
> tools etc., but also the political and organizational structure.
>
Debian is centralized, and I have seen no signs of it failing (other than
that they seem to be intent on rewriting dpkg to be more Nix-like instead
of just adopting Nix :-) ). I've heard some rumors about a NixOS social
contract, but nothing has materialized yet.

>
> I don't see much information about these points on your website (only
> technical stuff).
>

NixOS is pretty small now (only 21 people, according to
https://github.com/orgs/NixOS/people) so there hasn't been much need to
write it down.
If you look at contributors to the main Nix repo (
https://github.com/NixOS/nix/graphs/contributors), then it's even tinier;
Eelco wrote all the code (3400+ commits), and everyone else has contributed
to the Nix language or done building/porting fixes (209 commits, if I can
do addition correctly in my head). I've been thinking about writing some
core changes, and shlevy has started on a Haskell rewrite (
https://github.com/shlevy/hnix-store/), but in the meantime it is indeed a
one-man project with random contributions, following the "focus by limiting
core developers" criteria.

-- Mathnerd314
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20140728/cfae293f/attachment.html 


More information about the nix-dev mailing list