[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: debian mate

Michael Schmitt <tcwardrobe@gmail.com> writes:

> as I see you, as a member of ctte, are kind of in favour of MATE for
> jessie and not wheezy... :(

I have no opinion on MATE.  I personally switched from GNOME 2 to Xfce on
the one system where I use an integrated desktop when gnome-shell wouldn't
run due to the age of the graphics and the fallback for some reason didn't
function properly.  (This was *right* after it landed in unstable, and I
suspect this is a bug that was fixed eons ago.  I'd been meaning to try
Xfce since I make a personal policy of changing desktop environments every
few years -- I used to use GNUstep before GNOME 2 -- so I didn't bother to
pursue it further.)

I can say that, as a light GNOME user, switching to Xfce was trivial.  It
took me all of an hour.  But I can be a somewhat atypical user.

I haven't evaluated the quality of the MATE libraries or their
maintenance.  In general, I'm supportive of being discriminating about
what packages we include in the archive, since we're promising bug and
security support, but I'm also in favor of being inclusive and, in
general, welcoming packages for anything that people want to work on.

> Xfce or KDE might not be a catastrophe, but you must see the same issue
> as I see. You may not be as "paranoid" as I am, but we all know how
> "users" tend to be: annoying, complaining, crying. humans at its best!
> :) They have no right to complain, we all know that too, but does that
> prevent them from doing so? And the only perspective I can see there is
> trying to minimize the possible fallout...

I don't believe avoiding user complaints is a primary development goal for
Debian.  One of the great things about working on Debian is that we are
not a popularity-driven or market-share-driven project.  We want to do the
right thing for our users, but that isn't the same thing as avoiding

In the specific case of releases, we have a stark tradeoff when it comes
to freeze deadlines and release time frames.  On one hand, some users will
always want something that missed the freeze deadline.  On the other hand,
*all* of our users who want to run stable and not testing or unstable are
hurt by release delays.

The way that we, as a project, have chosen (after *much* discussion and
some experimentation with alternatives) to resolve this conflict is with a
freeze that's advertised well in advance, and a clear policy of what's
acceptable after that deadline.  There are other alternatives to managing
a release cycle.  They all have different problems.  This one seems to be
the best compromise.

> But I am not sure it would need that much more time.

It always takes more time to introduce things at the last minute.  That's
why we don't have this discussion every time, and instead have developed
some clear guidelines for how to make this decision.

We can literally have this discussion forever.  There's always just one
more thing that someone thinks is really important.  The advantage of
having clear guidelines is that we can say that it doesn't matter.  It
missed the freeze.  It's too late.  Best of luck next time.

It sounds harsh, and to some extent it *is* harsh, but I'm serious when I
say that without this sort of policy Debian will literally become

> But nice, we do agree that MATE is from a users point of view the BEST
> alternative for gnome2?

No, I have no opinion on that.  After all, my personal experience is that
Xfce is a fine alternative for GNOME 2 if one doesn't like GNOME 3 for
some reason.  :)

> And I know what you have in mind there, with "but down that path lies
> never releasing at all" it does just not fit here. Something like this
> does not happen regularly.

Quite to the contrary: something like this has happened, about this time
in the release process, in every release of Debian that had a freeze since
I started using the distribution.  :)

> That all said, I need to express my apologies for such long mails for an
> idea where almost anybody seems to be against it. :( I know for a fact
> what a wonderful operating system and community (with its humanly common
> exceptions *g*) Debian is and my contribution today (which some may
> consider to be annoying as hell) is this tl;dr-candidate for some
> readers. :) But imho the only right thing to do right now, even if it is
> so damn late... :/

Oh, I have no objection to you expressing your opinion!  By all means,
that's what the mailing list is for.  But a lot of past experience has
completely convinced me of the merits of hard release freezes.

We have backports.debian.org for those things that are just desperately
needed but didn't make it into the release.  If MATE makes it into
unstable/testing and is proven stable, backporting it to wheezy would be a
viable option.

Russ Allbery (rra@debian.org)               <http://www.eyrie.org/~eagle/>

Reply to: