Re: gnulib textual domain?

[Sergey Poznyakoff]

Paul Eggert <address@hidden> wrote:

> Could you please explain why this is needed?  For example, what
> programs would use the gnulib domain?

Not the programs, but translators. The gnulib domain should be a
compendium domain, like existing domains `error', `iso_3166' etc.

> How would they use it?

Sources from gnulib are incorporated into a lots of projects. This means
that translators of various packages have to do the same work over and
over again. For example, being a translator myself, I often encounter
the same msgids from gnulib modules in different projects. Of course I
have a compendium (a collection of transaltions I have already made)
that I use. Some other translators have them too, but in the result the
same msgid is often translated differently, simply because it is
handled by different people. In two words, the result is: (1) wasting
translators' time and (2) lack of consistency between different textual
domains.

Now, what will it look like if there exists gnulib textual domain?
Before a translator starts working on a project that uses some of gnulib
files, he first downloads gnulib.LANG.po file and incorporates it into
his project.po, using msgmerge. Thus he will get translated all messages
from gnulib and will be able to concentrate only the real translator's
work: translating the project-specific messages.

This is the way we use textual domains `error' and `iso_3166' that I
mentioned above.

> how would version mismatches be handled as gnulib messages are
> added/deleted/changed?

I can undertake keeping track of it. It will take a bit of my time,
but the result is worth it.

Regards,
Sergey