NOTES FROM A TRAIN

Or - what REALLY happened in Toronto?

One may wonder.
The day of the Decision was at hand. The cards were on the table; IPNG
was going to be selected!

The chair of the IPNG Work Area comes up, talks at great length,
finally presents the recommendation: SIPP with 128-bit addresses.

A wave of relief seems to sweep the room. Something like a feeling:
"Yes, they made a decision after all. Now let's get on with some REAL
work".

The meeting ended one hour early.

(One other note: The previous transition plan for switching from IPV4
to IP:NG, which focused heavily on translating routers, was deemed
"too complex", and will be replaced with a new transition plan, called
SST for "Simple SIPP Transition", which seems to focus heavily on
dual-stack hosts, a main, and much maligned, feature of the TUBA
transition plans. However, not all is clear. Watch this space.....)

(More notes on IP:NG: One expects that the *default* mode of operation
with IP:NG will be to use cryptographically signed packets - the
security drafts feature prominently in the IP:NG series.)

(It was expected that we had until about the year 2000 before IP:V4
would not provide universal connectivity; however, this was based on
data collected before the current publicity blitz and the Mosaic
takeoff)

Enough of the lower layers - that is about all I gathered from this
area.

E-MAIL
------
My main focus is, and will probably remain, E-mail.
Some things that are probably significant:

- The PEM (Privacy Enhanced Mail) group still doesn't have its act
together. The current draft of a MIME-compatible message format looks
implementable, but there are stil real problems. In another move, it
has been suggested to expand the PEM key-management messages so that
it can do a number of new things, including user identities expressed
as E-mail addresses, and the exchange of PEM keys!

- IMAP4, the mailbox access protocol, seems finished. Terry Gray
expressed the goal of this meeting as ensuring that nothing happened,
and it succeeded. (It was mainly concerned with disposal of comments
from the Area director). The main debate was whether an inactivity
timer should be 30 or 15 minutes; the debate occupied more time than
any of the suggested timer values.

- The EDI group is moving towards a conclusion about the technical
means of transferring EDI in MIME. The rest of the work is ensuring
that the EDI community understands the realities of operating in the
Internet environment, so they are writing an introduction document.
Dave Crocker, the group chair, claims that most of the time spent on
this group is two communities learning each others' languages.

- The advisory body on RFC 1327 met, and made a few recommendations:
  - There are a number of minor edits to 1327 that should go in.
  - One should use the work of the NOTARY group - but it is not
    certain whether one should use the full report format, or just
    rewrite the current report format into something "notary-like".
  - The body part conversion document should be kept separate, and
    needs updating.
Volunteers were found in the group to write updated documents, which
will be input to the normal IETF WG review process.

- NOTARY (Notifications) reached near closure on the format of
notifications. At the meeting, the issue of error codes was raised;
there is a need for a clear, machine-readable error code, and the
current draft uses the SMTP error code set, which is perceived by many
to be inadequate.
 Discussion quickly proved that the issue was neither clear nor
straightforward; one discussed various ways to get around the problem
without bogging down the real work; decision on this was not reached.
 The group also embraced a volunteer to write a document describing
how to do vacation notices in the same framework, and recommended that
the work on receipt notifications, which is much more controversial,
be handled by a separate working group at the next IETF.

- A new group called MAILEXT (Mail Extensions) was set up to handle
stuff that needs WG review, but seems too straightforward to be worth
setting up a separate WG for.
The items on the table today were:
 - SMTP extensions for streaming: considered uncontroversial.
 - SMTP extensions for binary/chunked transfer: Probably OK.
 - File-transfer body part mapping: Generated a discussion on the
   appropriateness of defining "nice-looking" headers with fixed,
   dubious meaning; deferred to the RFC1327 review WG (when
   established).
 - Language tags for use with MIME: considered uncontroversial.
 - Recommendation on returning 521 error code on port 25 of systems
   that will never handle mail: needs further discussion.
(I'm not sure I remember everything)
The decision on almost all documents was to make a discussion on the
group mailing list, and then submit for Proposed Standard.
The chair also promised to do some legwork to make sure "everyone"
knew about the new group; the first time I noticed the group was on
the printed IETF agenda!