Re: i18n, part 2

[Daniel Jacobowitz <drow at false dot org>]

On Tue, Dec 07, 2004 at 07:00:10AM +0200, Eli Zaretskii wrote:
> > Date: Mon, 6 Dec 2004 15:51:14 -0700
> > From: Kevin Buettner <kevinb@redhat.com>
> > 
> > The code in question is simply outputting a list of attributes
> > associated with a particular thread.  How should we make the necessary
> > context available to the translator without putting an undue burden
> > on GDB maintainers?
> 
> I suggested a way to do that.  I don't think it's putting an undue
> burden on us, but if someone comes with a better idea, I will gladly
> vote in favor.  Just let's not assume the problem does not exist.

I do think it's an undue burden, as I said earlier.  While the eight
existing strings in that function aren't too bad:
  (A) what happens if someone needs to add another option?
  (B) I would like to have a standard answer to this question for GDB. 
      For instance, infttrace.c:print_thread would need 128 (plus a
      dubious %s).  I hope we agree that's too many.  Of course, that
      seems to be a GDB-internal debugging aid; I wouldn't shed tears
      if it was never translated.

> > What do other projects do in this regard?
> 
> I don't have any statistics handy; the general request from
> translators is to give as much context as possible, as close to full
> phrases or sentences as possible.  I find myself in the need to read
> the sources while translating much more than I would like to, so I
> guess the maintainers don't do a very good job at that.  I'd like GDB
> to be better.
> 
> If we want more opinions and suggestions, we could take this to the TP
> mailing list.

I'm just concerned about the "as possible".  It may be possible, but it
isn't practical.  No, I don't have a suggestion to offer you; I don't
know enough about translation issues, but I do see that this change
would create a code maintenance issue.

-- 
Daniel Jacobowitz