A criticism I read of the OSI/ISO/CCITT process is once they define some
stuff for a particular area they start defining stuff for mildly related
areas in which they do not have expertise.
This discussion about linguistic details, language markings, language
representation, and how we can tag everything is very interesting.
However is it something which this group has a high degree of expertise in?
I have *some* knowledge in this area but cannot call it `expertise' as the
last course work was 5+ years ago.
No, probably not. Therefore my proposal is the following:
-- Leave the specialists' needs to the specialists. TEI (The
Text Encoding Initiative) seems to be the most promising
effort to cater for these needs. We should be prepared to
register content-types suitable for TEI-encoded text.
-- Make the existing, well-established ISO 639 standard usable
in the Content-Language: header field.
-- Prepare for accepting also the future improved three-letter
codes, when the new ISO standard eventuelly is adopted.
-- In the meantime, allow IANA registration of additional,
hopefully temporary language codes and variant codes, when
they are needed. This does _not_ imply that IETF has to
decide on the appropriateness of any individual code, and
even less that IETF has to design an all-embracing _system_
of language codes.