Sorry if I wasn't clear. Nathaniel's proposal sounds like the right one,
specifically because it can be ignored and most of the other ones can not.
Are you really going to ignore Nathaniel's example header when it
comes in?
Real-From: US-ASCII/base64: TmF0aGFuaWVsIFMuIEJvcmVuc3RlaW4=
Isn't this a case of handing your users incomprehensible junk, which
you strongly objected to on the issue of recursive encoding? Now just
ignoring mnemonic in headers is a reasonable (though not optimal) option.
But it just doesn't make sense for the Real-* headers.
Bob Smart
P.S. On Greg's point that "no one has objected to the Real headers". I
would point out that I said "Nor do I think the Real headers will
be regarded as satisfactory [by the Europeans et al.]". This was intended
to be a strong objection. However it is rather spooky conducting this
debate without a single European voice. Keld where are you...