Hi everybody, and sorry for the delay in commenting on this flame^W
thread. As a (lame) excuse this thread took me during a long-ish period
of (Debian-related) travels, that ended only now arriving at DebConf12.
On Thu, Jun 21, 2012 at 04:08:58PM -0400, Paul Tagliamonte wrote:
> As many of you know, we're tragically close to freeze. As a result, the
> usual process of voting, debating and picking a theme via consensus is
> not really possible. There was also a distinct lack of interest in such
> a process, as well.
>
> I've taken it upon myself to start getting desktop-base into shape for
> wheezy.
First of all let's observe that, looking strictly *from the point of
view of typical Debian procedures*, what Paul did is flawless. He picked
up the job on himself, coordinated with the desktop-base package
maintainers, and coordinated with the Release Team to ensure that an
important part of our forthcoming stable release was in good shape for
the freeze. For this work, and for the courage that it took to do that,
I think we should all be grateful to Paul: thanks!
Unfortunately, that is not enough to conclude that the whole "Debian
Wheezy artwork process" has worked at its best. Some of it is in need of
massive improvement, while others parts of it did work quite well.
Before looking at the details, let's ask ourselves while we (partially)
failed: because nobody picked up the job of organizing the selection
process. Yves-Alexis, who did the job for the Squeeze, made it very
clear that he wasn't willing to do that again. And that's a very good
start actually: stepping down considerately is very helpful for other
team members. But that's not enough to find someone else doing the
job. Which is why I've been calling repeatedly for volunteers, both on
this list and on debian-devel-announce, willing to do the job. Finding
no one. So I did try to cover it up myself, one step at a time, but at
each step I made it very clear that I didn't want to step on anybody's
toes and that I'd have preferred if some of the people active on this
list pick up the remaining tasks. (Still finding no one.)
So, what did work well? My answer: the call for help went well. We sent
out a PR on it and we got 24 submitted proposals for Wheezy. If the wiki
pages are to be trusted on this, we got only 4 for Squeeze. This is
massive improvement! No matter which decision process we choose and its
flaws, proposals are what allow to make a decision in the first place
and all artists who contributed a proposal have helped with that, even
if they haven't been chosen.
And what didn't work well? The timing and the (meta-)decision on how to
decide the artwork. Even if I didn't really want to do this myself, I
"gladly" take my share of responsibility on these two. I didn't really
realize (out of naivety) that a good deal of the theme should've been
ready for the Freeze. I was (implicitly) counting on freeze exceptions
and I didn't know theme-related exceptions had caused RC bugs in the
past.
Then there's the meta-decision. I did mention surveys as a possible
mechanism for choosing the theme, but actually never committed to it.
Originally, I wanted to discuss how to decide on this list. But more
recently, discussing with people who worked on this in the past, I got
convinced of doing it differently: nominating a group of people (a
"committee", if you want) that would've taken the decision. Given no one
else was doing the job, I thought it would've been legitimate to do so,
possibly (DPL) delegating the committee members to take the decision.
In fact, the day Paul picked up the task, I was looking for committee
members, starting first with informal /query-ies, as I often do for this
kind of stuff. I didn't have much luck with that either; of the people
who I contacted, only Yves-Alexis was willing to participate (thanks!),
although only as a fallback if I was short on candidates. I would've
escalated the call for team members to this list, but Paul stepped in
first.
In the end, that final timing part is no big deal, the problem is that
we were too late anyhow. Yes, that sucks, and I'm concerned because what
happened gets in the way of the ability of we, Debian "packagers", to
work with Debian "artists". I'm really sorry. But that's what happened.
I think it happened that way because we have been unable to find, for
more than 1 year, someone willing to coordinate the decision process.
That kind of "management" contributions are as useful as packaging
contributions and artistic contributions; we really need to find a way
to attract them.
Could we do better next time? Ulrich is right to have doubts, given he
has been dissatisfied with how the process went with Squeeze already.
My answer is that I've no idea. It could go better if: (1) we remember
this and (2) someone steps in and do the coordination work with
sufficient time advance. Either way, I think we should keep in mind what
went well (the call for artwork) and repeat it.
Thanks for the time you all put in making Wheezy a nice-looking release.
Cheers.
PS I think one specific criticism mentioned by Ulrich in this thread is
not particularly fair, the one about (paraphrasing) "I've submitted
on a proposal and I now feel I wasted my time". That specific problem
is not related to how we made the decision, it's a more general
problem of calling for several artworks and choosing only one of
them, be it by vote, committee, of benevolent dictator ruling. In
part, I disagree it's wasted time all together; as I mentioned above
we need different proposals to make a choice, that's useful
contribution already. On the other hand, we should (and could)
minimize the sensation of wasted time, by shipping alternative
proposals and make it technically easy to change theme, distro-wide,
in a given Debian release.
--
Stefano Zacchiroli zack@{upsilon.cc,pps.jussieu.fr,debian.org} . o .
Maître de conférences ...... http://upsilon.cc/zack ...... . . o
Debian Project Leader ....... @zack on identi.ca ....... o o o
« the first rule of tautology club is the first rule of tautology club »
Attachment:
signature.asc
Description: Digital signature