ISO 639 is too coarse to be so useful, anyway
From involvement in a number of UN projects, 639 is adequate to be
useful for most languages that are written and in which extensive
international communication occurs. There are certainly situations in
which something finer would be preferred, but they do not represent a
sufficient majority of cases for 639 to be "not useful".
In any event, the proposal to use 639 for language tagging is separate
from the proposal to *do* language tagging. If 639 were actually
inadequate, then we should either use something else or supplement it
with additional qualification.
There is no single spoken Chinese.
It seems to me that this would be an important issue only in text-to-
speech applications, and then only in those applications in which it was
important that the recipient "hear" the same Chinese as the sender.
Again, that represents some class of applications, but certainly not
all. And, while I can think of some examples, it is not clear to me
how often one would prefer text-to-speech to simply sending the speech
(as audio) if the sounds, rather than the characters, were relevant.
It is not clear to me that we must choose "no solution" (or a
solution that either creates a very large number of registrations and/or
gets us into trouble when 10646.2 is published) if there is no solution
that meets all possible needs.
john