[Fwd: [l10n-dev] l10n experiences]

[Simos Xenitellis]

To parakatw gramma einai synexeia tou Ougkrikou "peiramatos" tis
metafrashs tou OpenOffice sta Ougkrika mesa se 3 imeres.

Prwta episynaptw ta sxolia allwn mellwn tis listas l10n tou OpenOffice
kai epeita to gramma tis Ougkrikis omadas pou periexei perissoteres
plirofories gia to ergo tous.

Filika,
simos

A. =====================================================================

Από: 	Louis Suarez-Potts <louis at collab.net>
Απάντηση Σε: 	dev at l10n.openoffice.org
Προς: 	dev at l10n.openoffice.org
Αντίγραφο:: 	johnzero-ooffice at johnzero.hu
Θέμα: 	Re: [l10n-dev] Hungarian translation "party", questions
Ημερομηνία: 	06 Feb 2002 12:53:52 -0800	
Hi, Noll,
Congratulations!
I'd like to do an interview on you, for our "spotlight" section of the
homepage. The interview would be short (don't worry).
I'll email you directly later,

Louis Suarez-Potts
Community Manager

B. =====================================================================

Από: 	Eike Rathke <erl at sun.com>
Απάντηση Σε: 	dev at l10n.openoffice.org
Προς: 	dev at l10n.openoffice.org
Θέμα: 	Re: [l10n-dev] Hungarian translation "party", questions
Ημερομηνία: 	06 Feb 2002 11:08:06 +0100	
Hi Noll,

On Wed, Feb 06, 2002 at 01:39:24 +0100, Noll Janos wrote:

>  Maybe this is a bit offtopic, but I think you'd all be somewhat
> interested...

It's not offtopic at all, it's faaaaantaaastic! :-)

  Eike

-- 
 Eike Rathke, Software Engineer, OpenOffice.org Calc core developer.
 Number formatter bedevilled I18N transpositionizer.
 PGP key 0xDA620A55: 01 53 0B 6F E3 37 08 47  E2 A0 6A 63 C1 3E 8B 96
 .project: char*(*(*foo())[])()

C. =====================================================================

Από: 	Nils Fuhrmann <Nils.Fuhrmann at sun.com>
Απάντηση Σε: 	dev at l10n.openoffice.org
Προς: 	dev at l10n.openoffice.org
Θέμα: 	Re: [l10n-dev] Hungarian translation "party", questions
Ημερομηνία: 	06 Feb 2002 10:29:05 +0100	
Whow!!! Congratulations!!!

This is the fastest localization I ever heard of !!

Let's see how we can get the translations on the OOo page:
Some more comments and questions inside

>  The project began on last friday afternoon, and lasted until monday
> morning, and was mostly non-stop. About 12-16 people were in the
project
> room at a time, and aboul a 100 people collaborated over the net,
using
> our special PHP/CGI script. We translated the approximately 21000
strings,
> that the localize tool had extracted, in about 36 hours, then we
proceeded
> to building, testing and correcting the result.

Have you followed the instructions for the L10N-Framework to do the
extract?

>  How can we contribute the translation back to the CVS, to
Openoffice.org?

We have to discuss if we want to integrate all new localizations
directally into
the code (which makes the code less readable) or if we find an other
solution,
maybe a new module which includes the translated gsi files (extracted by
localize).

>  What more do we need to do? Where will we need to put it, who will we
> need to send the language-text file?

That's me. As I understood, you have the full localized gsi file at
hand. I stay
tunned to recieve the stuff.

One more question:

Have you already done a rebuild which leads to a runable installation
set? If
so, have you used the method described in the L10N-Framework
documentation?

Thanks and congratulations again

Nils

D. =====================================================================

Από: 	Sander Vesik <sv117949 at ireland.sun.com>
Απάντηση Σε: 	dev at l10n.openoffice.org
Προς: 	Noll Janos <johnzero-ooffice at johnzero.hu>
Αντίγραφο:: 	dev at l10n.openoffice.org, Barry Caplan <bcaplan at i18n.com>
Θέμα: 	Re: [l10n-dev] Hungarian translation "party", questions
Ημερομηνία: 	07 Feb 2002 11:09:13 +0000	
On Wed, 6 Feb 2002, Noll Janos wrote:

[snip]

>  By now, I think the whole thing could've been done with the converter
> scripts, as the PHP could've handled the input/output conversions.
> However, given the small amount of time we had, this didn't really
occur
> to me :-)
> 
>  By the way, the localize tool missed some strings, about a handful in
> the dbaccess module (don't know the name of the resource file, but it
> was one file). By using "tags", it was easy to spot these strings.

Wow! This is really cool system 8-)

<plug type="shameless">Have you considered donating & hosting it at
openoffice.org?</plug>

=====================================================================

-----Προωθημένο Μήνυμα-----

From: jerry <jerry at gnu.hu>
To: dev at l10n.openoffice.org
Subject: [l10n-dev] l10n experiences
Date: 11 Feb 2002 21:37:23 +0100

Istvan,
 If you can wait a little longer, we hope to release our codebase
so that you too can have a localized OO in the same effective
manner. Also, we didn't build a glossary beforehand: it was
just created "on-the-fly". There was a responsible who took
care of updating a webpage where people could find non-trivial
or disputed translations. It worked well. We can still have a
postgresql dump for a full glossary like the ones you can find
here... we are planning to do it after the whole stuff was
lectured, revised. We never needed a full 13 thousand wordlist
though. IRC and a mailinglist came handy enough.
The tag system also helped to find the exact context of a string:
"left" can have postfixes in hungarian in certain circumstances -
this was the only way to find out which kind of "left" we needed.
 If you set up a webserver like ours, anyone can join - and within
time, the strings will be taken care of by volunteers.
 During our debug process, we set up a bugzilla reporting system
where users could report the TAG of the mistranslated, mistyped, or
simply too long (for a button for example) strings. (note: our 4
char random tags collide sometime, but didn't drive to any
confusion...)

 One little note if you too plan a 3 day marathon: be aware that
when backing up the database, it can be locked: it can be really
annoying at "standard cron time" at about 4am. At first we didn't
know what was happening ;-)

BTW: the windows install kit is exceedingly difficult to build. We
haven't managed yet: we only have linux binaries and are working on
BSD. The web directives on w32 are more than inaccurate: for example
we had to downgrade java to be able to proceed. Simply put: w32 src,
in our case, does not build out-of-the box. Is there an updated HOWTO
somewhere?

The ISPELL for hungarian is currently a bit buggy: it does not
recognize accentuated capital letters nor "composed" words which are
common in hungarian...

bye
peter