>sam. 10 janv. 2026 at 13:02, Gabriel Wicki <[email protected]> wrote:

> Hi there, fellow Guix!
>
> I recently picked up the great opportunity of a (paid) gig for the Free
> Silicon Foundation (F-Si).  The F-Si received a grant to push free
> software in the EDA (Electronic Design Automation) space, where
> proprietary software with license fees up to $1M/year/user dominate.

Great news, indeed ! Congratulations.

Let me play the devil’s advocate for a moment: remember that the most
usual suspects in EDA are almost free of charge for research labs and
the like, see Europractice for example. And the order of complexity they
deal with is orders of magnitude beyond what we could expect with free
tools. This is the starting point.

> Since there is plenty of free EDA software, the F-Si figured the main
> goals were to map, package and advocate already available free software
> solutions—and that's how they found me!

Making EDA available and keeping it in a good shape is the first step
and one of the goals of the guix electronics team, for sure. But there
are many other aspects to improve from our side:

- let it know: checkout any of our ghdl related packages, you’ll see
  they advertise installing with Guix; this is not the case of most
  packages (which do advertise using Nix)

- architecture support: see yosys build status other than x86_64,
  similar occurs for the rest of our EDA packages

- features, ci, containers, reproducibility, profiles, dependencies,
  installing, generations ... all usual suspects in daily problems which
  Guix fixes for good

- replace existing [1] attempts to EDA bundled suites, can you imagine
  how easy would it be with Guix to provide images with everything
  already inside, ready to run your favourite pipeline ?

- and many others ...

All in all, this comes to time and manpower. To me, the way to go is
using Guix as foreign, no one will replace their beloved OS just to get
Kicad installed. So simplifying as much as possible the process is a
must.

>  1. Should we introduce our efforts to the general public through a blog
>  post addressing both the Guix and the F-Si crowd?  I have a draft ready
>  that I will soon send through the list.  I am not sure if our blog is
>  really the right medium for that, since it is mostly a promotion of
>  Guix for electronics designers, but maybe this still fits?

No idea of what a gig is, but if you count on the support of a big
organisation, you may take the opportunity by getting inspiration on
what HPC is doing around Guix: dedicated channels for EDA tools not in
Guix (yes, they exist), including IP blocks, a dedicated web domain,
your introductory post, technical articles, a good documentation for
people willing to try, a dedicated mail list, a means to exchange, the
unavoidable wiki ... the idea is to create a community around using Guix
for electronics design, broadly, and the people is already there [2].

>  3. Could we eventually maybe build bundles of software (through
>  manifests) on our CI infrastructure?  So people anxious to install and
>  use Guix on their machines could still get a good use of our efforts?
>  This seems like an (almost) free addition once packages are built
>  already.  WDYT?

In my experience, building substitutes when you’re outside of the main
Guix umbrella is the main limitation to gathering a community, remember
that.

My two cents,

[1] https://github.com/YosysHQ/oss-cad-suite-build
[2] https://fossi-foundation.org/orconf/2025

--
Cayetano Santos
.
gpg: CCB8 1842 F9D7 058E CD67 377A BF5C DF4D F6BF 6682
key: meta.sr.ht/~csantosb.pgp

Attachment: signature.asc
Description: PGP signature

Reply via email to