ietf-nntp IETF 41 summary -- Comments Welcome

[Stan Barber]

There are three working group documents currently in development.
Comments about these documents plus a new draft submitted by Brian
Court were the core of the scheduled discussions at the meeting of
the NNTP Extensions working group at IETF 41.

Stan Barber outlined the agenda for the meeting and summarized the
status of the various drafts that would be discussed during the
meeting. The "common practices" draft is in the 12th release and
there have been few comments on it since the last meeting. Stan
expects that this draft will be the basis for the final version
that will be submitted to IESG. The "rfc977bis" document is in the
9th release. There have been many comments on this draft since its
last release, but most of them have centered on AUTHINFO. More
about this later in these minutes. There has been an update to the
NNTP SRCH extension draft. There have been very few comments on
the draft since its previous release. Finally, Brian Court submitted
a new extensions draft to allow a dynamic feed adjustments.

There were very few comments from the group about the need to make
adjustments to the "common practices" documents. Stan proposed that
the draft be revised to include some examples and a table of contents
and presented for a WG last call before next IETF. Assuming there
are no last minute problems, the draft will be moved forward for
consideration by the IESG for publication as an informational RFC.

During the discussion about the "rfc977bis" draft, there was a
lengthy discussion about excluding the AUTHINFO command from this
document and doing it in a separate extensions document. This was
proposed as a tactical move to help get the "rfc977bis" document
wrapped up and publishable. There was some concern there would be
problems with this since it would mean that the base protocol would
not include any form of authentication/identification command.
However, other folks noted that the original NNTP spec didn't have
this and most folks don't depend on these features. So, AUTHINFO
will be removed from this draft and put into an extensions document
of its own. There was a volunteer to create this new document.
Unfortunately, I didn't capture the name of that individual.

There was also significant discussion about how to represent the
"end" of a posting that has a body that is all binary. The current
spec says that the end of posting is signified by a period on a
line by itself. Does this mean that the end of a body is CRLF.CRLF
or just .CRLF. Folks pointed out that NNTP was not really designed
to support binary postings. I don't know that we really came to a
consensus on this. I do know that we want to avoid specifying
anything about the format of the article. That's really the purpose
of RFC 1036 and any update to that document.

In discussions about the SRCH extension, the group noted that the
keyword in the draft was still SEARCH not SRCH. The authors noted
that was an error and would fix that. There was a significant
discussion about the internationalization of the search terms. The
authors intended to specify that the SRCH term was to be encoded
in UTF-8. It was argued that there might be a need to support other
encodings since the article body may support other encodings. The
authors said they would consider this.

Finally, Brian Court presented more about his draft about a process
he named "dynafeed". The core idea is to permit a downstream to
inform the upstream of changes in a feed it is receiving from the
upstream as it is occurring. While there is merit in the idea, the
direction of the communication as expressed in the draft is not
supported by the NNTP model. The downstream is supposed to respond
with response codes, not commands. Brian will take his idea and
revise it for future consideration.

Respectfully submitted, STAN BARBER