After a couple of days' work to push it up the last 10%, the British English translation of GNOME 2.22 is complete, barring any further string changes, and assuming the changes I committed to GTK+-2.12 actually turn up in the translation statistics (which, worryingly, they haven't done so far). It's verging on being late, I know, but it's better than nothing.
Hopefully this is OK for my first attempt at translating GNOME.
With bkor's help, and after a dummy commit to GTK+, the translation is now officially at 100%.
Damn, I was wanting to do that. Ach well, at least I won't have to put up with 'z's and no 'u's now. Thanks!
Congratulations on the achievement, shame a task like this that one might thinks was easy is as complicated as it is.
Now if only all the other en locales like Ireland, Jamaica, and anywhere else that didn't want American spellings could also take advantage of this without needing to duplicate the effort like the Canadian English team does*.
(* Okay so maybe there are a few spelling differences but the differences between American and British English are small and the differences between British English and Canadian English - in Gnome at least - tend to be even smaller. A little more rephrasing of the American and the differences could be even smaller.)
I'd love to hear more about your translation work. Does the process of translation to British English prove any easier or more automatable as a dialect of English than translation to an entirely foreign language? Do you start from the C or en_US translation? Do you use any automated substitutions for the most notable spelling differences? What kinds of non-spelling differences come up in software translations?
I can't really compare the process to that of translating to a completely foreign language, but I presume it's a lot easier; most of the translation can actually be automated. For completely untranslated modules, I start from the C locale and use en_GB.pl (http://svn.gnome.org/viewvc/gnome-i18n/trunk/en_GB/en_GB.pl?view=markup) to convert it to en_GB, then check through it manually. For partially-translated modules, I just run `intltool-update en_GB` then use gtranslator or gedit to fix whatever strings need attention.
The only real non-spelling differences which have come up are inconsistencies in – or missing – punctuation; I've removed a fair few extraneous commas, added some hyphens in, and in some cases rearranged strings so they make more sense. I'm not sure if this is allowed, but I think it is, and nobody's told me off for it yet. 🙂
It sounds like many of your non-spelling changes would apply to the original text. You should push those changes there rather than in a translation.
I do plan to eventually, but I couldn't do during the string freeze. Thanks for reminding me. 🙂