L10N migration to transifex.net

Yuri Chornoivan yurchor at ukr.net
Mon Feb 21 11:26:54 UTC 2011

> Le 21/02/2011 08:12, Yuri Chornoivan a ?crit :
> > написане Mon, 21 Feb 2011 08:10:55 +0200, Ricky Zhou
> > <ricky at fedoraproject.org>:
> >
> >> Sorry to say this, but it's not quite a vote of confidence yet I'm
> >> still trying to get more of a understanding on the reasons that some
> >> people don't want to move to transifex.net.
> >
> > From the point of view of a translator (that is my private point of view):
> >
> > 1) transifex.net is as trashy as Launchpad was not in its wildest days.
> > Anyone can register the project with no obligation to its translators at
> > all. Noone looks at the licensing of these translations. Launchpad at
> > least does not allow projects of "just for me" translations. If this
> > practice on transifex.net will be continued, the server of the project
> > become very overloaded by the trash.
> >
> > 2) Many developers seems do not care at all (or do not know anything)
> > about updating the templates of their projects (packagekit,
> > shared-mime-info, quassel, freeipa(?), et al.). I believe that current
> > scheme of 1.0+ does not work and should be rearranged. It is very unusual
> > for developers and translators.
> >
> > 3) Interface of 1.0+ is very crowded and overloaded. There is no way to
> > switch back to classic interface. In Tx 0.9 it is suffice to push one
> > button to do anything. In 1.0 you have to push two or more buttons to do
> > the same thing. Search function does not work for localized messages.
> > Translation memory does not work (anyway Lokalize do it in magnitudes
> > better, but the fact is not very promising).
> >
> > 4) polib (the engine of new transifex) suffers from severe bugs due to
> > implementation of some Python string functions. These bugs can lead to
> > impossibility of localized Fedora installation.
> >
> > 5) Current t.fp.o installation "just works" regardless to the frustration
> > that we have in the past.
> >
> > There is no time enough to test everything thoroughly before Lovelock will
> > be released. The bugs in Fedora 15 can severely damage GNOME 3.0
> > reputation.
> >
> > That is why I think that transition (even transition to 1.0 in t.fp.o)
> > should be postponed to past Lovelock period.
> >
> > Best regards,
> > Yuri
> > Ukrainian translator
> I'm very surprised about your reaction.
> The l10n project ahs been requesting for help for about a year now. Now 
> that some solution has been found and decision taken, you seems to wake 
> up to disagree.
> No perfect solution exists, what do you suggest according to YOUR time 
> and availability to make it work?
> Regards,
> Thomas


Please do not get me wrong, but the only thing I wanted to say: "It was the wrong time to change the horses". If it was two (or even one) month ago, I would strongly say "Yes!!!" (just for myself) and did my best to iron out the glitches.

I do not wake up just now. I had a lecture during the session on this question, so could not attend it. 

All these things were decided in three days (maybe I missed something?). Some people might not even recognized this (see commits to t.fp.o). There were many discussions last year in this list, but the change in a first day of freeze was just like lightning strike or Micronokia deal (fast and right).

Now I can only say that we have no time before release. I invested MY time into Fedora translation and if now it is just a matter of leaders decision to make a move that will ruin parts of my work... Well, I will do my best to solve the issues, but cannot guarantee that the small team like ours can invest enough of OURS time to make things run as they should in three weeks that left (we did not even see the new documentation).

That's all.

More information about the trans mailing list