Language codes (was Re: Serbian (sr) language translation team:maintainer unresponsive)

[Christian Rose <menthos menthos com>]

tis 2003-04-08 klockan 18.20 skrev Charles Voelger:
> > > I may be wrong, but this is how I understand the system should work.
> > > maybe you could instead use a language code for yekavian seberian, if
> > > there is no ISO two-letter code it offers me the oppurtunity to ask any
> > > one who may know, what should one do in a situation where the language
> > > doesn't have an ISO code? 
> > 
> > One requests for an ISO code to be assigned for that language. See
> > http://lcweb.loc.gov/standards/iso639-2/.
> 
> That was interresting reading.  Is there a reason that we use the two
> letter codes for locales, and not the three letter?  From reading that
> site it seems the three letter would be more appropriate for
> translations, is this simply a historical issue, or are there practical
> reasons for the two letter codes being used?

I guess both. The three letter codes are a newer standard, and when
translations began once upon a time I guess only two letter codes were
available. But that is pure speculation on my part.

Fact is that renaming every single po file (and expect glibc,
distributors and end users to rename their locale names on their systems
to fit) would be a huge amount of work for very little actual gain.
Also, two letter codes are nice and one less character to type. :-)

So I guess that's why we currently use two-letter codes where available
(and use three-letter codes where not). In fact we currently don't have
translations for any language that doesn't have a two-letter code, but
that might change soon. I heard Kenneth was investigating whether there
would be interest in doing Low Saxon translations (three-letter language
code "nds", no two-letter code).


Christian