I, too, dislike dropping all mention of MNEMONIC.
I thought we had agreed in Santa Fe to a definition of MNEMONIC that
would satisfy everybody (the use of MNEMONIC to mean "mnemonic embedded
in USASCII text, with + as the leadin character", with strings like
"mnemonic-iso8859-29" to be reserved for use "between consenting adults").
It is going to be used, so dropping it from the standard is a BAD mistake,
I think.
On the subject of binding with RFC-HDRS, the text on the bottom of
page 1 is good enough for me - pushing both through the standards
process is politics, not RFC text.
Harald Tveit Alvestrand