Re: latest draft - content-transfer-encoding

I've been following the 10646 and UTF discussions quietly
because, like Ran, I think that the thing is in enough turmoil
that "Standard first, MIME-plan later" is the right model.
But I was struck by Olle's excellent (IMHO) analysis of 
penalty factors.  If one looks at that analysis, and the 
conclusions he draws about alternatives to UTF for people using
non-ASCII alphabets, one may have a strong argument for a "forget
10646, just design and use a good 2022 profile".  :-(
    john