Comments on new draft (text version):
1. I'm not like the new syntax for the mailserver access-type. Why can't
you make it another parameter, like "command=" or something. Do you
really expect this to be a multi-line thing?
2. On the richtext converter in Appendix D. You might add a line at the
end that says on "UNIX" systems, one would likely pipe the output of
this program to "fmt".
3. On page 40, para (1), "star with" -> "start with"
4. I agree with your suggestion that the "Q" encoding in RFC-HDRS be
normalized.
5. I think the references (e.g., RFC-HDRS) can stay in.
All in all, well done. Can we close now?
/mtr