(iso639.551) Many languages - how to code?

John Cowan jcowan@reutershealth.com
Sun, 5 May 2002 18:50:26 -0400 (EDT)


John Clews scripsit:

> 3. A requirement for a IANA tag for language X is made. No ISO 639
>    code exists, and Michael Everson, as IANA Language Tag Reviewer
>    allocates a yet different code [i-def]. This might be because
>    there is a more urgent need for this code (indeed perhaps a set of
>    codes, as in the recent request for a great many North American
>    languages) than is likely to emerge in sufficient time from the
>    deliberations of the ISO 639 Joint Advisory Committee.

Fair enough.  But given the recent mapping from SIL to ISO 639-2 tags,
I think that the RFC 3066 i- tag is effectively deprecated for new
uses; I would expect any new RFC 3066 tag to have a primary tag that
is ISO.   So at worst we would have a conflict like this:

Suppose (quia absurdum est) that a need were found for allocating a 
tag for Yahi, an extinct Central Yana language of California.  I am
not certain, but I suspect that NAI is the currently relevant ISO 639-2
tag.  So the IETF would allocate a tag like nai-yahi, whereas the ISO JAC
could allocate the tag YAH.  This situation is really no worse than
that faced by a pure ISO 639-2 agent who tags Yahi text or audio as
NAI, only to be told later that YAH is now the correct code.  It will
be necessary for such an agent to be resilient in such a case anyhow.
(Yahi was the native language of Ishi, the last "wild" Indian in the
United States.)

> Deprecation by the standards developer certainly doesn't stop
> deprecated codes being used, and of course there may be various
> consequences as a result.

No.  But whenever there are more general and less general codes,
an equivalent result can occur.  The alternative is to throw up our
hands and say "no code for this language specifically, ergo no code
at all", which is why RFC 1766 was developed to supplement ISO 639(-1),
which had no more general codes.

-- 
John Cowan <jcowan@reutershealth.com>     http://www.reutershealth.com
I amar prestar aen, han mathon ne nen,    http://www.ccil.org/~cowan
han mathon ne chae, a han noston ne 'wilith.  --Galadriel, _LOTR:FOTR_