/var/log/mrpouit

Amélioration de la configuration des dispositifs de pointage

Ajout d'un éditeur pour les associations de fichiers

(French: désolé, un billet en anglais de temps en temps ça ne fait pas de mal, je le traduirai à l'occasion.)

A few weeks ago, I received a mail about a community survey. Inside, I could read that "[Jono is] in the process of trying to understand the needs, opportunities, and challenges of our community members [...]".

"Maybe they'll be able to show interesting patterns in the results, and make good use of them for the UDS". That's what I thought (I know, I'm still a bit naive and stupid, even after 4+ years in the ubuntu community).

The survey closed on Oct 22th, and on Oct 24th, Jono published his report . Obviously, that was not a good sign (that's too quick), but I read it anyway. Let's write it: if this kind of report is what is used by Canonical to try to improve various areas of the community, I totally understand why many people feel it's getting worse and worse.

In this reportthing, there are 28 pages of, mmh, analysis? (let's call it like that for now). To sum up, you can read that X% of the respondents come from Europe, Y% work at Canonical, Z% work on packaging, etc. etc., for 27 pages! Then you have the conclusion, based on the summary from above. Then you have all answers from the open text boxes in the survey. And, then, you have http://www.jonobacon.org/2011/10/24/ubuntu-community-survey-next-steps-leadership/ because of course leadership issues are the biggest issues in Ubuntu...

Yes, you read it right: this report doesn't contain any analysis at all, it's actually a summary of the raw results. Worse than that, Jono only reported global data ("26% of the respondents do not see themselves as leaders, so let's write something about that, so it will look like I care, and thanks to that I won't speak about the real problems."). The community isn't one and unique, and many different people participated in the survey ; there were many areas involved (cf. q3): accessibility, packaging, documentation, translations, support, loco, etc.

And yet, Jono thinks of them as a whole: do packagers really have the same feeling as translators, or testers, or documentation writers about their work not being worthwhile and appreciated? Of course not!

That's what makes this survey worthless, even misleading: not a single time in this report I was able to see a correlation between data! To give a few possible examples: what a packager wants to know, is whether other packagers have issues in their interactions with other ubuntu teams. What a MOTU wants to know, is whether she's the only one to feel that she's wasting her time because nobody gives value to universe anymore. What a Canonical employee wants to know is whether other employees also feels like they are discriminated in the community.

One more example: As a packager, I don't care if "the general retention rate that we appear to have with more than half (57%) of the respondents contributing for over four years". Many of the packagers who were around when I became MOTU 4 years ago have either disappeared, become inactive, gone on illimited vacations, or reduced their contributions a lot. Furthermore, what's the retention rate for testers compared to bug triagers? Who get demotivated the faster, documentation writers, or leaders?

What was the point of asking some many questions in the survey otherwise, if you don't correlate the responses? What was the point of publishing such an incomplete report instead of the raw data?

"bring back the community leaders from 2 or 3 years ago. they had a lot of respect and they worked hard. jono preaches about burnout as does the community, but nothing is done within the community to help combat it." is one of the responses you can find in the appendix (sorry, I didn't write this one ;-). And this is not new: when I was still involved in loco teams a few years ago, I heard it several times. I even heard myself advising several times possible contributors to not get involved in Ubuntu because of that.

Whether this is done willingly, by laziness, or ignorance of how to exploit survey results, the very ones who harm some parts of the Ubuntu community are the community managers themselves. And I can understand why some Canonical employees feel they are discriminated: when you hear all day Canonical community managers saying that everything is fine everywhere and there's no issue (easy, they don't ask people, e.g. derivatives about their relations with the Canonical/Ubuntu Desktop Team), you tend to be a bit nervous after working for three months on a issue (that doesn't exist, right?), and maybe you tend to be not so nice with the people involved[*].

We might need better community managers, if we want to solve the issues. Ones who do not think writing everyday everywhere that unity and ayatana are so great is part of their job. Ones who see the people in the community for what they really are, try to understand what they feel, and ones who do not only see the community through their own vision (very convenient for not being aware of issues!).

[*] I often speak with one Canonical developer of all the issues I met when interacting with some of his colleagues, and I'm sure it's sometimes annoying for him, even if he is part of the ones who care most and try not to break derivatives too much. I'm sorry, I hope he's not the one who wrote "Being discriminated against for working for Canonical, I contributed just as much before I was employed here and was employed based on my contributions." about what's demotivate him. ;-)

Quelques extraits des réponses des développeurs de Thunar au dernier troll en date sur xfce4-dev (traduction libre de votre serviteur, à la va-vite, donc à faire hurler les puristes, mais j'aime bien quand les puristes hurlent) :Donc j'espère que ça vous éclaire sur l'absence d'onglets dans Thunar alors même qu'ils existent dans d'autres gestionnaires de fichiers depuis plus de 5 ans.

(oui, il manque gvfs-backends installé par défaut sous Xubuntu oneiric, c'est le genre d'oubli qui gâche tout…)

Quelques nombres extraits de mes build logs (avant leur suppression) pour les deux dernières versions d'Ubuntu. Ils représentent le temps passé à construire, dans un chroot propre (avec sbuild), les paquets binaires (*.deb) à partir des paquets source (*.dsc et *.gz). Ça ne prend pas en compte le temps à faire/mettre à jour les paquets source, créer/appliquer d'éventuels patches, tester les paquets, etc.
Mes seules contributions récentes (au projet Ubuntu) étant sur Xubuntu, ces chiffres concernent probablement des paquets Xfce uniquement :

• natty : 65h 18m 22s
• oneiric : 30h 0m 33s

(Je n'ai pas de chiffres pour les versions précédentes, car je n'utilisais pas sbuild.)