Re: open translations database

["Kenneth Christiansen" <kenneth gnu org>]

> "Peteris Krisjanis" <peterisk@apollo.lv> writes:
> 
> > I think biggest problem for translators is that there's no place in web
> > with:
> > *translation specifics - HOWTOs, FAQ, etc.;
> > *translation formats, projects, etc. (GNU gettext, etc.);
> > *information about translation programs;
> > *other kind of information and possibility to host team page and project;
> > (of coarse, they can use sourceforge.net, but there it can be more specific)

We have to gather some of all the information available and put it on 
developer.gnome.org i18n section. Who wants to be responsible for this?

> > And in the end, about 'bad' or 'good' translations - IMO, relatively good
> > translation is when in fits in such factors:
> > *it's understandable to 80-90% of users of that language (balance between
> > tech. specific and user-friendly terms)
> > *it's uses specifics of languages, e.g. uses proper apex, etc.
> > *it's done almost in time (for every release) - but it's last factor, not so
> > important;
> 
> You're completely right.  Remember, GNOME is a free project (freedom).
> Pressing project members (the hackers) will make things worse.  The
> gettext manual has nice sections on this issue.
This is not about pressing anyone, but about finding the right workflow,
and putting the ease of the hackers. Maybe the update.pl should have an
--release option, that will generate a statusreport, and send it to the
gnome-i18n@gnome.org with information about a release coming up. Or maybe
I should just make a small newrel.pl script instead. What do you think 
about this. This should be quite simple, and make it a lot easier for 
hackers to get by

> You've to be very careful if you think it's the best to change the
> current workflow and the current tools chain (gettext collection and PO
> file format).

I will rather extent the gettext tools right now, that replace them. 
but first it's really importain to know what we really need! Like why
impliment glade support, if glade will be changed to use the new Gnome
i18n XML format. (I'll probably impliment that to glade when time
allows - actually now to break things, then after 1.4)
 
> As a proof of concept I'd like to see a prototype based on XML -- but
> please, stay away from proprietary Java implementations ;) For the time
> being Java is no fun on GNU/Linux.

Well, gcj is pretty much fun ;)

Regards, Kenneth