On 07/20/2009 01:25 AM, Keith Bowes wrote:
>> My problem with that is that we already have a localization mechanism,
>> that takes into account that GtkBuilder is a Gtk specific implementation
>> detail.
>>
>> In short it introduce a problem for the 2 other platforms.
>>
>
> I don't see how. It only affects the GTK-specific files anyway. It's
> not like I'm ripping code out of the Windows and Macintosh versions.
>
> Granted, it does introduce a level of redundant messages to the .po
And that's exactly the problem. We had so far avoided to duplicate the
localization, with a very good reason. Duplication is extra work for
unpaid volunteer that do it. The less, the better.
> files (for example, GTK+ uses underscores instead of ampersands for
> keyboard shortcuts), but that's no worse than the exact same messages
> that only differ in case now, which occurs in the .po files currently.
> Just remove the fuzzy flag, replace the character(s), and go about
> your business.
We had a workaround for that. Granted that we could have delt with it
better, I don't think that *duplicating* messages is the best approach.
> This was part of the project description from the beginning. I put it
> off until last being how nightmarish intltool is. I really didn't see
> a point in it, since the dialogs were being translated fine after the
> change to gettext, but my mentor mentioned how the way it's being done
> now is suboptimal. I believe his exact word was "hellish."
And having 3 different translations for 3 different platform isn't
suboptimal? I guess that the problem was not thought far enough, and
interested person haven't been really asked. Not to claim the Mac port,
and I do think that I'm the one here who know it best, along with the
Gtk code.
Hub
Received on Mon Jul 20 08:06:09 2009
This archive was generated by hypermail 2.1.8 : Mon Jul 20 2009 - 08:06:09 CEST