Archive for July, 2007

2.20’s internationalization; guadec results.

Wednesday, July 25th, 2007

maintainers

don’t forget to add new files with translatable strings to POTFILES.in/POTFILES.skip, otherwise those new strings will not be caught. quickly check http://l10n.gnome.org/module/$MODULENAME to get a list of those files (they are listed at the top of the “UI translations” section under “Notices” – if there are none, you’re a good maintainer(TM) and deserve a cookie.

translators

GNOME’s translators are very creative with regard to working around bad strings. however, it seems that they are way too friendly by not filing bug reports against modules with hard-to-translate strings. why?
are you keen on trying to translate “Read” (“has been read”? or “something to read”?), “Profile” (“a profile”? or “to profile”?) or “%s at %s on behalf of %s” properly by try&error guessing, or do you love spending time on reading the source code?

you’ve got better things to spend your time on! there aren’t many bug reports on translation issues in gnome bugzilla, which surprised me. so please file bug reports against the affected modules and add the “L10N” keyword, and if a string definitely has to be rephrased then please also add the “string” keyword to your report. help the developer to understand your problem, by linking to a good explanation, e.g. if a comment would help to understand how to translate a string, add a link explaining how to use comments, if a sentence is splitted up, add a link to your bug report to the explanation why this is bad behaviour. it only takes you a minute and will help everybody to get a better localized desktop.

everybody

if you like your favourite desktop and want to give some of the love back, why not getting involved by helping your translation team?

just take a look at the webpage and write an email to your translation team leader that you are interested in getting involved (if you do not receive an answer to your email within two weeks, please complain to gnome-i18n mailing list and CC the team leader on your email, because this should not happen.).

we have a webpage that provides wonderful statistics about which modules are in need of an update – choose your language, click on “GNOME 2.20 (development)” and then click on the download to disc icon of a module. then take gtranslator, kbabel, poedit or your favourite text editor to make the translation perfect again, then let the team leader upload your file, that’s all. easy, eh? :-)

languages in need of more translators

for my guadec talk (page 15f.), i compared the translation stats of 2.14 to those of 2.18 to find out which languages probably need more manpower. beside of course those languages that have always been on a low level, the following languages with more than 50% ui translation have lost quite a lot:
nepali (-20%), albanian (-14%), indonesian (-12%), czech and croatian (-11%), romanian and nynorsk (-10%).

also, i wonder whether those translation teams that have 99% ui translation (and therefore obviously enough human ressources) but less than 3% documentation translation (arabic, dzonghka, macedonian, hungarian, catalan, finnish, lithuanian, danish, japanese, vietnamese) could perhaps also translate a few docs for 2.20? we all know that a fully translated ui is much more important than having translated docs, but perhaps, perhaps some people can spend an additional little amount of time on translating the doc of their favourite application (again: if new translators are interested – of course feel free to join and help, see above).

general disclaimer: if you do not like translating, but want to help making GNOME even better: there are also many other ways how you can get involved.

post-guadec

“and now for something completely different.” looks like some of the issues that i described in my guadec talk are on a good way now:

  • willem is interested in providing better analysis tools for bugzilla, something that could make it easier to analyze/identify bottlenecks and the quality of our modules and services. looking forward to it.
  • if i remember the last guadec evening correctly (we were all so damn busy with fixing the showstopper bug 455415), don told me that somebody talked to him about working on providing a better tracking of our documentation status. i don’t know if it’s the same thing as gil described herevertimus, used by the frenchies, looks quite advanced, compared to the stuff used by the dutchies (i don’t remember the exact url, this one here is for the docs) or germans.
    yes, one evening at the etap, we introduced the systems used by the translation teams to each other. that’s why guadec is so cool.
  • i’d like to thank danilo for addressing our gtp bottleneck.

also, i ran into willie walker (working on orca, gnome’s screenreader) on his last evening at guadec, who paid me a nice compliment. thanks, man!

i’m a pretty happy camper currently.

at the bug front, fronting.

Sunday, July 22nd, 2007
  • congratulations to the tracker crew – bug 403752 has now become number one in the all time statistics. though it has been fixed in svn for nearly 3 months now (2007-04-30), the maintainer(s) is unable to either release a new version (0.6.0) or to isolate a patch that could be backported to the distros to stop this flood. mart (who is one of the brave guys triaging that one) told me that he has increased his gnome bugzilla points from 7 to 17 just because of marking duplicates for this one. by average, the bugsquad closes 7 dups of this one per day (!).
    jamie, maintaining your project does not only mean to write new code all the time, it also means to take care of those bugsquaders that support you. i would probably be less pissed if i had not been told for five times in the last three months so far that a new release will be done “within the next two weeks”/”at the next weekend”/whatever. last answer was “this weekend”, we will see what happens (or not). at least it’s highly unprofessional behaviour and not acceptable.
  • libsexy’s bug 354559 is still second worst in the duplicates statistics of non-fixed gnome issues. after waiting for six months, we now have a nice trace since one month. i’ve send two emails to the maintainers so far, no reaction. david, christian, please fix it and attach a patch for backporting it to the distros. thanks.
  • guadec has been very productive yesterday evening with regard to eliminating whiskey and south american drinks. please continue.

NP: Lupe Fiasco: Kick, Push

gnome, where’s my monitoring at?

Thursday, July 19th, 2007

guadec, birmingham.

currently at birmingham for one week to attend guadec, the gnome conference. my plane from cologne to birmingham had a delay of more than four hours, and i dislike waiting for a long time at airports (i know that already), but at least i was officially allowed to use the power sockets somewhere near some stairs, directly in front of the security check-in, so i could spend the time working.

when i arrived in the uk, i was first of all reminded of the fact that the uk is not part of the schengen agreement and that i had to queue up to show my passport when entering the country. it was way after midnight so there was no public transportation available anymore. i shared a taxi with two other folks that had been on the same plane, and got off at my hotel (note to google maps: your birmingham map is wrong, please put the hotel location where it actually belongs to. yes, i want mapy.cz for whole europe!). the rooms are fine and it’s pretty nice to sit in the entrance area/breakfast room, combine adapters with multiplicators with multiplicators with multiplicators to have at least 15 computers on one electricity line and finally break two of the power sockets at the wall. one elevator was also out of service, i wonder when we will have succeeded in burning down the hotel.

one receptionist asked me last night what all those people are about, and that all those computers look so “intellectual” (err… it’s the very first time i heard that term about geeky folks). i tried to provide a very basic explanation of what gnome is about (world domination?), but the word “code” did not tell her anything and even the comparison with that other “operating system” out there did not work out (i asked if she would run windooze on the family computer, but she could not remember). so i tried the olpc-some-of-us-are-working-on-a-better-world explanation and since then i think she’s a bit more confident with all those geeks sitting here at late night, blocking laying the tables for breakfast. ;-)

the conference so far has been cool, happy to see old friends again in real and to run into people that you had talked to on irc but never seen before (for example the south american bugsquaders and also some sun china a11y folks). the venue seems to be some kind of a concert hall (there are also small chambers where people can exercise playing the piano) and a few associated rooms and buildings next to the town hall. big advantage: carpet nearly everywhere on the corridors (gimme the real, dirrty hacker feeling!). the number of power sockets is acceptable, and the wifi can be stable sometimes (depends on where you are, there are rooms at the venue where it’s quite alright).

the city? to me it seems that birmingham did not have a reproducible urban development concept within the last decades, the town center and the pedestrian zone look okay though. you can find a camera nearly everywhere, for criminal prevention and for your own safety, of course. hmm. i have another definition of freedom.

my talk.

my talk today was fun and my first guadec talk ever. after sayamindu worked around the classic resolution problems of my laptop (thanks!), i talked to about 25 people in the audience. my talk was about using statistics to see the state of some fields of gnome. gnome already has some nice stats in the fields of translation (l10n.gnome.org), documentation and bugzilla, but i miss(ed) some statistics so i partially tried to gather the data, e.g. changes of translation rates or maintainer changes within the last year, amount of and response times for freeze break requests, missing announcements and so on. i also showed a bit of bugzilla’s auto-reject feature and how many reports have been rejected so far, so people get a basic impression what’s possible and especially what’s not possible currently. though i was mostly personally interested in finding some things out, i guess that the sheets should be interesting for translators and the bugsquad, and perhaps also for documentators and the rest of the release-team:

Where’s my monitoring at? (PDF, ~1,5MB)

i’m still missing some conclusions from the data that i have gathered, but we will hopefully reach some progress on those fields. i am looking forward to some related talks (for example fer’s talk, and the patchsquad bof).

and no, i have neither sit next to claudio, nor did he eat one of my peanut cookies. good guy!, as well as vincent who uploaded the sheets of my talk to his webspace… dĕkuju a dobrou noc!