Mailinglist Archive: opensuse-translation (117 mails)

< Previous Next >
Re: [opensuse-translation] l10n process improvement
  • From: Jean Cayron <jean.cayron@xxxxxxxxx>
  • Date: Wed, 11 Nov 2009 08:26:14 +0100
  • Message-id: <c62ec50d0911102326s50898a5et8acac869c4673bbd@xxxxxxxxxxxxxx>
Hi everyone,

We're of course discussing and it's only ideas.

2009/11/10 Kálmán Kéménczy <kkemenczy@xxxxxxxxxxxx>:
Hi Jean,

On Tue, Nov 10, 2009 at 1:24 PM, Jean Cayron <jean.cayron@xxxxxxxxx> wrote:
2009/11/7 Kálmán Kéménczy <kkemenczy@xxxxxxxxxxxx>

idea 5
We have to build l10n volunteers respect and community. I am not sure
why we should deliver languages where translations are under 75%. (at
Mozilla this is 100%). We have to contact to the translators and find
out what happens and maybe we have find more/new volunteers.

That's not very respectfull for small or minority languages that
strive to have the most translated they can, dealing with openSUSE,
Gnome, KDE, Mozilla...

Try to find new volounteers does not always
succeed. So what? Throw the language away? What's the point to have
Yast 100% translated (even some very specific and technician parts) if
half of the desktop is untranslated?

First of all, thanks for your feedback. I am just talking about ideas
here. There are no decisions (and I am not the one who could or want
decide anything alone).

About the translation percent. I am not talking about YaST, I am
talking about the distribution we release. And now I see I was not
clear. The localization is the hardest part when we are talking about
a whole distribution. How can I say to a school, government or users
that openSUSE is localized, when it's not. And we have no idea at all
the l10n status of any release. We really don't know what is missing
or what we have. The stat page is represent 5% (or less) of the
distribution and this is not good overview.

In that case, the "Warning" made about Yast could be used for the
whole thing but we should keep "developping" languages. It should be
clear in the promotion (wiki, sale DVD...) what are the "full
experience localised" languages. But there, the problem is how to
"rate" it.
Rating on a % of KDE, Gnome and openSUSE has no sens, as I said, want
there are loads of small programs not widely used on both desktops.
Essentials for KDE makes more sense. But again, openSUSE doesn't ship
unofficial translations of KDE, so it's already the case. Other key
applications are OpenOffice and Firefox. For these we can say: OK it's
officially localised. For Gnome I don't have a clue of how to deal
with it.

Or it can be done on the packages shipped on the OSS repo.

And what about GNU tools ;-)

Idea:
Yast: X % yes/no
lcn: X % yes/no
Gnome: ???? yes/no
KDE: official KDE released language yes/no
Firefox: official Firefox released language y/n
OpenOffice: official released language y/n
other important projects, application?

How to make stats about all these projects? That's a lot of work to do
for all languages, indeed and I'm not sure that scripty can do it.

And there we don't speek about quality, only abou %age. I remembered
having heard a sad story about a Mandriva based distro "Caxa Magica"
or something similar that had a contract with Portuguese government
for school computers. In their distro there was GCompris and it was
translated in Portuguese by one of his creator, from Portuguese origin
but living in France. Some translations were a bit "unusual" and the
teachers made a fuzz about the "badly translated educational program",
the press got involved and so on, despite the GCompris community
brought a quality update on that translations within the week.
It ended up a bit sadly with the educational minister explaining how
to remove the "evil" application without mentioning that there was an
update available...

There is no good way, except feed-back from a release to another, to rate that.

But again, if something is done, I'm for the warning, not the withdraw
from the translations.

I think we have to find out when we could say something is localized.
Which components are key and where should we improve. Maybe we have to
find new contributors on different languages where there was no commit
for years.

On KDE, the language is released if the essentials are met:
desktop_kdelibs.po : 75%
desktop_l10n.po : 75%
kdebase : 75%
kdelibs4.po : 90%

These files/folders are about 20% of the whole KDE project but they
are the core. The other parts are not evaluated for release.

On GNOME, they don't care.

According to the stats, following your idea would make 23 languages
only. There are 60 languages translated on openSUSE.
http://i18n.opensuse.org/stats/trunk/toplist.php

Speaking of my case, I'm very proud to have managed to come from 45%
to 57% from 11.1 to 11.2 alone, despite I try to find other
translators.

We should and we are proud of this effort as well.

And I work also on KDE and Firefox (with a very few other
people). Should I give up or spend all my hometime translating it? Or
stop translating KDE to be sure openSUSE gets 75%?

No, we should build community for each languages. The one man show not
works well in the community development.
I know but more you have speakers, more you have volunteers. And more
you have systems/apps localised, more you attract people. It's a bit
the case for me now, some people start to propose their help because
they've been attracted with a product.

Mozilla is far smaller in number of strings than openSUSE.
Translations on openSUSE are far more work intensive than, for
example, Mozilla or KDE want it's more technical and it includes the
documentation in the GUI. They're several POT with weird name with not
any explanation of what it's for on the web. Should I translate it
badly to get the %age or should I leave it now untranslated?

imho, you are already know the answer for that. Bad translations are
worst than no translations.

I can understand than some languages around the 20% or so are a bit
"short" but for the others, please think about the consequences. If an
"essential" requirement is set, it must be defined wisely and the POT
files have to be cleaned up from some outdated pot's (kinternet for
example...).

We are here to talk about it. And again, I am just one small piece of
the list and I will not and can't judge anything. I just would like a
conversations between us, how can we leverage the processes.

thanks
kalman

No personal problem with you Kalman, if you pass and visit me in
Belgium, I'll offer you a free homemade beer with pleasure ;-).

Regards,

Jean
--
To unsubscribe, e-mail: opensuse-translation+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse-translation+help@xxxxxxxxxxxx

< Previous Next >
Follow Ups