Re: Guix wiki

[zimoun]

Hi Tobias, all,

On Tue, 11 Jan 2022 at 16:30, Tobias Geerinckx-Rice <me@tobias.gr> wrote:

> Guix cares about community very much.  Many of us care very little
> for wikis, having seen how they attract much outdated and
> incorrect information and spam.  Wikis are high-maintenance.

I agree with these words…

> I must say, the proponents of wikis have done an exceptionally
> poor job of representing them.  Yet we're supposed to be convinced
> that we're missing countless valuable contributions by people who
> can't be bothered to send a mail.

…but as I tried to explain elsewhere [1], different interpretations for
the aim of this “Cookbook” seems floating around.

 - cathedral, as it is today
 - bazaar, as a wiki could be

Obviously, it is possible to participate to the cathedral; clone, tweak,
drop a patch, done.  All by email.  But somehow, it could appears a bit
rigid at first.  Imagine, (generic) you are fighting with new paradigm,
your distro is not working as expected, you do not understand what’s
going on, after some reading here or there, and several attempts, you
succeed!  Now, one advice you found was not working exactly and you
adapted it a bit.  I am doubtful that you will read how to contribute
and finally send a patch.  I think,

    $ git log --format="%an" -- doc/guix-cookbook.texi \
          | sort | uniq -c | sort -n

makes my point. ;-)

Do not take me wrong, I am also on this side (accurate and maintained
documentation via review; something that might last) but it appears that
some people would also like half-baked poorly-maintained documentation
as a wiki could provide.

There is no conflict, IMHO, because the goals are different.  For
instance, EmacsWiki [2] saved me a lot of time, although information
there is often half-baked or poorly-maintained and back in time, it
helped me to then jump to the Emacs Manual, Emacs Lisp Manual and others.
Similarly for ArchWiki [3].

It is not because I would not personally contribute in feeding this that
this kind of material cannot be worth. :-)

Therefore, IMHO, the question is becoming: how to gather this kind of
material?  Because we have to admit that the rigid Cookbook via email
patches of texi file is a limited success for this bazaar goal.  An
official wiki is one idea; it raises where hosts it?  How maintains the
infra?  etc.

I proposed [1] a Planet, other proposed ’Awesome list’, and from my
experience, it could be a first step to collect.  Again, where hosts it?
How maintain it?

Other said, do we, as a project, officially support such documentation?
For sure, with one big warning that the material can be outdated, even
wrong.  Or do we consider that we have already enough in our plate and
many place can host unofficial such documentation?

I am fine with both. :-)

Last, for sure, I am not convinced that officially hosting a wiki would
automagically feed it after the buzz of novelty and the hosting cost is
not negligible.  That’s why I think an intermediary as Planet or Awesome
list is better; for the current resource we have at hand.

1: 
<https://yhetil.org/guix/CAJ3okZ20FdcnJ244RRdnPumBnu=moRLqPeHHxBzfUHQNZytrCg@mail.gmail.com>
2: <https://www.emacswiki.org/>
3: <https://wiki.archlinux.org/>


Cheers,
simon