Rick's comments notwithstanding, UNICODE in its present form doesn't
fully and unambiguously support all languages (Thai and Vietnamese are
among those not sufficiently dealt with). I think that the ISO DIS
10646 is closer than UNICODE to supporting all languages at present,
though it isn't perfect either.
It might be the case that at some future point that some such
"universal" standard will be developed, but we aren't there yet
and so support in the draft RFC to mark the character encoding
standard used is necessary.