Re: adding support for missing languages

[Jordi Mas]

Hello Andrew,

Thanks Karl.  This always bugs me.  We currently have
absurdities such as la-IT for Rennaiscance Latin.
What has the country code for Italy got to do with a
variety of a language belonging to a historic period?

Well there's no country code for "rennaissance" you
may say.  Exactly.  No country to me means don't put
a country - not pick whatever country.

We currently have locales with the language-country pair, such as EN_IE (English Eire) and locales with no country such EO (for Esperanto).

For languages that really do not have cultural conventions is better not to specify any. If you think that 'Rennaiscance Latin' should not have a country attached (that I agree) we should change the locale.

However, you should take into account that people that already has document marked as la-IT will not be recognised as it if we change the locale, at least if we do not keep both internally.

Anyway I've wanted a real solution to this for ages
and a proper discussion about it so we can design it.
Please check out my oft-advertised bug report:
http://bugzilla.abisource.com/show_bug.cgi?id=3227
"Need to extend language tags".
I think we need to invent a "language object" which
can hold more and more precise info than the mere ISO
tags.  We should use the objects exclusively and only
(lossily) convert them to the ISO tags where necessary
for foreign file formats.  Our own file format should
of course have a non-lossy equivalent.
There I said it (:

I COMPLETELY agree that we should move from the current two letter language into a better system. For example, we currently cannot 'support' (put Alan's right word here :) ) Asturian (bable) because it does not has a two letter code, and many other examples that you point out in your bug report.

As everyone else, I also have an idea how this move can be done. My suggestion is to use support three charters for the language name, wherever is possible we use the 639-1 and if there is not language code we use 639-2(B). For example, for Spanish locale in Spain we use CA-ES and for asturian we use AST-ES. This is the system that for example Mozilla.

Regarding cases such as Pitjantjatjara, Warlpiri and Arrernte, they are not common. You can always play with the two letter country iso code to diferenciate between them, if one day we have the need, that I really do not think so.

This is just my opinion!

Thanks Andrew,

--

Jordi Mas
http://www.softcatala.org