Re: Relational database-support for glossary and translation memory

[Christian Rose <menthos menthos com>]

Veronica Loell wrote:
> I'm planning to implement a relational database backed tool to aid
> translation. The idea is to utilise all the information and knowledge
> that exists in the i18n part of the open source community.

Great! :-)


> A couple important points:
> 1. Unicode.

Traditionally, all language teams choose themselves the appropriate
character set for their translations. This information is also specified
in the individual po file headers, and the tools AFAIK have to respect
that.

The exception is .desktop files, which don't contain any easily visible
information about the character set used. I think we all agreed upon
enforcing UTF-8 encoding for .desktop files in GNOME 2.0.

Using UTF-8 encoding in the database itself is probably a good idea, as
long as there can be database output (with tools) in other character
sets.


> 2. Integration with translation tools that are used, ie for updates
> and extraction of information.

I think integration with gettext tools (po tools) is essential. This can
probably be done with extensions to xml-i18n-tools.
Both updating of the database from po files and the extraction of
information in the database into po files would be nice to have.


> 3. The ability to easily be able to share information among the
> community of translators.

Being able to store and update the database in the GNOME CVS is probably
a good strategy for GNOME translators. I think this makes an XML
solution interesting.


> I am relatively new to this community and it would be very helpful
> with feedback about both the idea of this project as well as the
> tools that would be preferable to integrate.
> In order to do the initial design I need to get a lot more knowledge
> of the basic tools used today, so pointers to that kind of information
> is greatly appreciated.
> 
> Veronica Loell

http://www.gnu.org/manual/gettext/html_mono/gettext.html is probably a
good resource. It describes the po format and the use of gettext tools
for translation of programs.

I don't know of any URL for xml-i18n-tools though. The authors are
Kenneth Christiansen, Darin Adler and Maciej Stachowiak. I'm cc:ing this
to them since they may have some pointers (Kenneth is already on this
list).


Christian