Re: Fwd: Re: Bug#727708: init system coupling etc.

[Olav Vitters <olav vitters nl>]

 [ pretty tired so this message is harsher than intended ]

On Sun, Feb 16, 2014 at 02:46:02PM +0100, Frederic Peters wrote:
> Hey Sri, nice of you to join the discussion,
>
> > I think eventually though we are going to have to monitor and see where the
> > discussion is going.  Debian is an important distro for us and we should
> > make sure that we continue to be the default.  I think we need to make sure
> > that people understand why we are depending on logind and why it is a bad
> > idea to have to support every init out there.  Now if the Debian team wants
> > to take on that task, so be it.  But this doesn't seem like the right thing
> > to do.  It's best to try to get people on the same level and convince
> > Debian that logind is the right technology and that we don't want to
> > support anaything else at least on Linux kernel based distros.
>
> At least Olav and I were following the debate, and Olav engaged in the
> discussion several times. From mails, IRC, and private dicussions I am
> of the opinion that on this matter the situation is quite clear for
> the Debian GNOME team, it's just that Debian sometimes is a place of
> conservatism, and arrangements may have to be made.

We should make clear what our position is. Reasonable demands are fine.
Going to us because init systems are lacking is plain weird. I was very
friendly in my reply, but it needs to be said that going to GNOME for
init system development is weird.

>   As an illustration, you could consult https://bugs.debian.org/681834
>   and https://bugs.debian.org/688772, they are about the dependency of
>   GNOME on NetworkManager.
>
> About GNOME as the default for Debian, the potential default changed
> to XFCE in November, and this will be re-evaluated in August (Debian
> git web interface seems down, but http://archive.is/erpyk archives
> that).  If you have any idea about things GNOME could do to help, it's
> now a good time, just before 3.12, do not hesitate to bring bug
> numbers to our attention list. (in the perspective of the Debian
> discussion, I believe a11y support and classic mode are important)

Rather arbitrary reasoning that you give here. XFCE was and does lag
behind in a11y support. Classic mode is something new and we were
already the default. I don't think these things matter that much, it is
more about gut feelings and if the full feature set works across all
architectures and kernels. E.g. rather have desktop which pretends to
work, than one which is clear and explains where things will break down.

Debian has a very conservative feel to it. IMO nicer if we are NOT the
default, and just have a spin. Then we could have something which shows
GNOME in a way that we know will work. Not something which is changed
for sake of theoretical technical superiority of "portability" and
"choice", while in practice just being buggy. I really like GNOME to
work on *BSD, but the way Debian seems to go after that seems wrong, it
will cause bugs and missing features that will make the experience bad.

The link you give state "How well the UI works for both new and existing
users.". I think that leaves out "as guessed by a few people".

Making "Software" work on Debian seems great. IMO we should ignore the
weird stuff and just focus on making GNOME shine (irrespective of any
distribution). With weird I mean that they're not trying to cooperate,
but trying to blackmail to remain the default or in Debian.

I've spoken to 3 Debian GNOME package maintainers during FOSDEM. They
said they don't speak up much, they prefer just doing their thing and
making GNOME work. My intention to Debian is that they should focus on
providing a good experience. Irrespective if we are a default or not.
Positive cooperation, because if being a default results is why they
cooperate or demand things, then that is the wrong reason.

I'm totally cool with proposing to GNOME foundation to e.g. raise some
money or something to assist Debian in achieving logind like interfaces
on *BSD, etc (if technically possible). But that should be based on a
request for assistance and first we should make clear what forms of
cooperation (communications) are acceptable. I think it would be cool
anyway to enhance *BSD on behalf of GNOME. But at the moment it seems
better to continue doing that on our own. Too much negativity and
forcing going on. I *really* dislike it.

> > > Also I trust the Debian GNOME team with their decisions, and they know
> > > how to reach us if necessary.
> >
> > I prefer to work with them on this, if only to know how the conversation is
> > going.  Debconf will be in Portland this year and so I will be there.  It
> > will be important to know what's going on before going amongst people there.
>
> Great to have you there.  There's a debian-gtk-gnome mailing list but
> work and discussion mostly happen on their IRC channel, it's
> #debian-gnome on OFTC.

Now using Polari. It is a bit simple atm :-P

-- 
Regards,
Olav