ON LANGUAGE NAMES /// RE: Results of Duplicate Busters

[Frank Ellermann]

Doug Ewell wrote:

>> I'd guess the general understanding of language subtags derived
>> from ISO 639 is that they are symbolic, coded elements
[...]
> I've always worked under that assumption -- that language
> subtags represent the languages themselves, unlike ISO 639 code
> elements, which represent the names of languages.

There is a subtle difference, the names for say "de" and "en" are
stable, but the languages are moving targets, without a standard.

Even if you'd consider "Duden" or "OED" as some kind of standard,
which edition corresponds to "de" or "en" ?  If it is simpy the
latest edition, are all older "de" and "en" tags obsolete, do we 
need an expiration or timestamp mechanism ?

Typically we don't care, and codes for the names are good enough. 

Sometimes we care, using year numbers in (or for) variants.  Or
using different subtags for historic variants of what could be
considered as the same language, "frm"/"fr", "en"/"ang", etc.

> The same goes for region subtags and ISO 3166, and script
> subtags and ISO 15924.

Also moving targets, especially the regions, solved by RFC 4646.

For scripts we had no serious problems yet, but I still refuse
to consider IPA-gibberish as proper Latin.  The folks trying
to register IPA as script variant like Latf / Latg had a point,
it is a border case, admittedly not quite like Cher.

 Frank