Re: Guix and the developer ecosystem

[Simon Tournier <zimon.toutoune@gmail.com>]

Hi,

Thanks for your feedback.

On Wed, 26 Jul 2023 at 19:29, Distopico <distopico@riseup.net> wrote:

> So, in many cases, I haven't been able to fully integrate Guix into my
> development workflow. I had to look for alternative ways outside Guix,
> like Nix.
>
> I appreciate the simplicity of Guix, but let's say that Nix has a
> developer-oriented approach and has become very popular among
> programmers. Many projects now include default configurations for Nix in
> their repositories.

Hum, it seems an another issue about bootstrapping. ;-)

Well, on one hand, Guix is missing some person-power.  And on the other
hand, this allows to easily contribute, somehow.

Considering Haskell, the upstream community is putting their hands on
Nix.  For instance, people involved in GHC release management provide
some Nix stuff [1].  Well, if you follow the Haskell community, maybe
you know that some companies [2,3] are providing resources for making a
smooth and friendly Haskell development experience relying on Nix.

Similarly, if you are in academia, the experience using Guix for your
computations should be friendlier.  For example, working with R using
Guix is just smooth.  Or how easy is to apply package transformations in
order to benchmark some HPC stuff.  That’s because some research
institutes [4] are providing resources for making a smooth and friendly
experience.

Do not take me wrong, Guix is not focused on academia issues and Nix on
development ones.  Both are free software and the limitation is people
motivation, somehow.  Personally, I mainly use Guix for doing
computations in scientific context so most of my motivation is to
improve this area.  Improvements and the way to cover your needs is by
discussing your troubles and then by trying to fix them, i.e., by
sharing your hacks – even tiny ones.  Maybe another fellow hacker will
find that cool and they will be motivated for collaborating.  Thanks to
this collective sharing, we end up with tools that cover our needs, more
or less. :-)


1: https://gitlab.haskell.org/bgamari/ghcs-nix
2: https://well-typed.com/
3: https://www.tweag.io/
4: https://hpc.guix.info/


> Another issue is that if I wanted to bring Guix into the development
> workflow in a team, there would be the limitation of the OS. While I
> promote free software in working groups, not everyone uses the same OS -
> some use GNU/Linux, some use Mac, etc. I think this is also part of the
> reason why Nix has succeeded in development environments.

Somehow, I agree that having Guix running on other OSes than GNU/Linux
will be helpful in attracting users.  The archive provides such
attempts, for instance:

        Guix on macOS
        from Chris Marusich <cmmarusich@gmail.com>
        Wed, 11 Oct 2017 20:29:57 -0700
        https://yhetil.org/guix/87bmldavre.fsf@gmail.com

First, please note that it’s far to be trivial to port Guix on any other
OSes than GNU/Linux.  Then, from my understanding, the blocking point
is: you need to cheat using a very large binary seed or workarounds are
somehow required and need to be applied against the core C library.
Therefore, the intersection between all these deep technical
difficulties and skilled people able to deal with them is currently
empty.  Mainly because Guix is rooted with GNU principles and motivation
is not fungible.

That’s said, I know people that are running Guix on MacOS using Rosetta
and friends.  Well, that’s not the right place to discuss that because
we have to stay focus on free software. :-)


> 1. 
>                                      or something similar to the
> overwrite feature in Nix flake?

Could you explain what Nix flake is using Guix terms?


Cheers,
simon