Re: Guix package incubator (later a channel)

[Pjotr Prins <pjotr.public12@thebird.nl>]

On Tue, Feb 07, 2017 at 08:38:58PM +0100, Ricardo Wurmus wrote:
> This makes me wonder if the presumed benefits of an incubator could
> actually be realised.  I would like to advise against recommending an
> “incubator” procedure like this as an official alternative to
> submissions to the mailing list.

I don't have the intention of making the incubator a formal part of
the Guix package submission system, so there is no reason to recommend
it. It is also clear to me that the current workflow works for the
current and experienced package maintainers.  The current ML-based and
bug-tracking system remains the authoritative one. You don't have to
be involved with the incubator.

It is also clear to me that the current system has a high barrier to
entry and actively prevents people from starting up and even
continuing. So, I want an alternative. 

The incubator invites anyone to submit patches in their own way on
their own trees. These contributors are encouraged to work towards a
final patch which will be submitted to the ML - that is when you take
over.  Exactly what an incubator does: small chicks get to be full
grown and egg laying chicken (what happens to the males you don't want
to know).  Maybe some of us will never outgrow being chicks. It will
be a learning experience for me too.

The current state of affairs is that people are working off the main
ML and packages never make it into Guix. I am a case in point. I want
more people to track my 100+ packages. With the incubator I hope to
increase visibility of people dabbling in Guix. It is not a large step
since I have been working this way with a number of other people for
some time. You can check the logs of genenetwork/guix-bioinformatics
on github. Some packages made it into main line, including the ldc D
compiler.

So, it serves a purpose. It does not require your involvement. It may
lead to more people contributing. In the worst case, the incubator
fails as an experiment. Main choice now is that we move away from
github - and I think that is a good idea. If the GNU git repo could be
forked on savannah we would do that - but that is no option. So, I am
looking for an alternative. 

Pj.