my blog

Sunday 19 November 2006

MyGNOME bits #1

I've seen some Debian people doing DWN-like posts and I thought it'd be a nice way to let people know about what I (or others) have been doing lately. Okay, this will probably not be in the real DWN format, but it doesn't matter that much. And don't think the next posts like this one will be as lengthy: I'm cheating and talking about old stuff in this one :-)

Foundation elections

Looks like I'm running again this year. At first I didn't really want to do it, but after talking with a few people and after seeing there were not enough candidates, I sent my candidacy. Of course, many people nominated themselves within the last hour before the deadline :-)

Instead of having big plans for next year, I've thought it might be better to focus on little things that are helpful. This is probably what I should have done this year. It's also nice to see some new candidates running: it makes me feel that you can get involved in GNOME and feel part of the community quite fast (and we can still improve a lot here)

Membership Committee

Good to see that Barış completed the new membership management system. This is something I really wanted to finish before I left the committee last January, but I couldn't achieve it. Now applications can be processed quite easily and are all tracked with rt3. I've seen the committee in action (you can look at the membership-committee archives) and it's quite impressive.


I've been cleaning the gnome-desktop bugs lately. Now most of the bugs are related to the fact that we're using GnomeCanvas (I can't wait for a canvas in GTK+ :-)). I've also deprecated GnomeDItemEdit since nobody uses this anymore and it's just not GNOME-y.

The next big thing is killing most of the icons we ship in this tarball. I've discussed this a bit with Andreas (who nearly died of horror when he looked at the icons). The hard issue here is that people are using those icons for their launchers and we don't want to break their configuration (you know, compatibility, etc.), so we can't simply remove them. Having symlinks to icons in the gnome icon theme or a gnome-compatibility module sounds like two possible solutions.

I guess I can also ask for help to our translators to fix this bug.


Some cool action happening here. First, Tom Tromey started triaging bugs for gnome-session. What most people probably don't know is that Tom is one of the first guy who worked on gnome-session in 1998. In the last few weeks, he's done a wonderful job with the bugs, and even started writing some patches. I seriously hope he'll maintain this module soon (he needs help from the sysadmin team to update his ssh key).

There's also Dan's work on session management. The goal is to get session-related code right for 2.20. Rock!

As for me, I've been reviewing and committing patches. I also started rewriting the session properties capplet in python. In its current incarnation, the capplet looks bad:

  • most items in the "Session Options" doesn't mean anything for most users
  • I don't know of anybody using multiple sessions
  • the "Current Session" tab should probably be a view of gnome-system-monitor

So we're left with the "Startup Programs" tab, which needs a lot of work anyway. The thing is, a capplet with only this doesn't look really nice, and I didn't feel motivated enough to finish it. I should still do it. specs

In the last few months, we've been slowly improving some specs, The desktop entry spec saw the most work, and it's starting to look really good now. The autostart spec has seen discussions (thanks Dan!), but we need to agree on some stuff before updating it. We should also probably look at the menu spec which could do with some love, I suppose.

I also intend to update desktop-file-utils since it's probably not doing a perfect job at following the spec (especially since we updated the spec ;-)).

CVS cruft

The sysadmins moved 200 old modules to the archives repository. I'm sure there are still many modules in the 600+ ones we have in CVS that could be moved there. In fact, I even started a list of modules that we could move there, but it would need confirmation from each maintainer. If someone wants to help here (by sending a mail to the relevant maintainers), just contact me. We'll then be able to enjoy a CVS with less cruft. And maybe, one day, a faster conversion to SVN :-)

Release Team

Awesome mail from Andre (why aren't you on Planet GNOME???). This is really something where we sucked in the past, and I'm glad that our newest release team member is doing the bug nag work. It didn't get a lot of replies, though.

I've also been amazed that releasing the latest GNOME versions has been simpler than ever. I guess there are two reasons for this: our external dependencies list and the build brigade. We should give more love to Frederic for tracking HEAD buildability.


Wow. Long time since the last one. We really need to get another one out soon. The one about the About Dialogs looks good, but is blocked by this bug. Or maybe we can simply try to finish the gnome-doc-utils migration?

JDLL 2006

I didn't blog about this, but GNOME-FR was present at the JDLL. It's always fun to be there. We had a nice booth thanks to our volunteers and the GNOME Event Box. We got some ideas on how to improve the booth and ideas of things that should happen in GNOME (probably more on this later). I'm a bit sad that we didn't think about having a wall of love, though.


I'm now trying to clean my GNOME INBOX. There are many mails there that probably need answering, and I feel bad about this. If you're waiting for an answer, don't hesitate to send a ping!

Friday 17 November 2006

We Are Sorry to Inform You ...

Dans le monde de la recherche, il y a une chose qui est pour moi assez frustrante : pour exister, un chercheur doit publier. Beaucoup. Pas forcément avec la qualité qu'il souhaiterait. Cela colle peu avec mes tendances de perfectionniste pinailleur, d'où ma frustration.

Et pourtant, parfois, on parvient à écrire quelque chose dont on est satisfait (et non, je ne prétends pas avoir atteint ce stade ;-)). Reste à franchir la barrière de la publication, et donc la barrière des relecteurs. Voir des critiques (que l'on juge) infondées ou causées par une non-compréhension n'est pas vraiment agréable. Après un refus, et quand on voit les articles acceptés, on se demande parfois où est la logique dans la sélection, ou plutôt on se demande si la logique est réellement la qualité de la recherche.

Un petit article illustre parfaitement ce problème : We Are Sorry to Inform You ... de Simone Santini. Il s'agit en réalité des commentaires de relecteurs de six articles, des commentaires totalement négatifs. Pour des articles qui sont considérés maintenant comme des classiques ou de grandes avancées.

Pour preuve, je laisse chacun méditer sur cette conclusion d'un des commentaires cités :

Publishing this would waste valuable paper: Should it be published, I am as sure it will go uncited and unnoticed as I am confident that, 30 years from now, the goto will still be alive and well and used as widely as it is today.
Confidential comments to the editor: The author should withdraw the paper and submit it someplace where it will not be peer reviewed. A letter to the editor would be a perfect choice: Nobody will notice it there!

(L'article en question est devenu une lettre bien connue, avec un changement de titre non choisi par l'auteur.)

by Vincent