2821bis, 2822upd, email-arch

Hi, the 2821bis and 2822upd drafts are "almost" ready, and clearly
better than the older RFCs.  Similarly the email-arch draft is 
"almost" ready.  A few open points are however critical, maybe
controversial, and getting them wrong would be dangerous.

Mail and SMTP is as important as HTTP, and the 2616bis folks got
a proper WG to sort our their nits.  Mail and SMTP deserve the
same care.

* 2822upd:  The NO-WS-CTL issues are critical, they do not only
  affect 2821bis, and checking all details is a difficult task.
* Email-arch:  Recent drafts such as MINGER and TBR need a term
  "border" or "boundary" MTA, and email-arch should explain what
  that is as common reference.
* "Bounce address":   IMO this is no *concept*, it is only an
  obvious *function* of the "envelope sender" concept formerly
  (STD 10) known as "reverse path".  It needs some fine tuning
  of 2821bis (mostly done) and email-arch to prsent this as
  unique idea, essential for SMTP, like a Message-ID is essential
  for NNTP.
* IPv6:  We should at least discuss Doug's proposal to get rid
  of the address fallback for IPv6.  Maybe the 2821 "bug" of not 
  mentioning AAAA can be turned into a 2821bis "feature".

 Frank