ietf-nntp My notes from the NNTP WG meeting at the 37th IETF

[Stan Barber <[e-mail]>]

These are not the official notes of this meeting, just mine. However, I would
appreacte any corrections or comments. Please post them to the list.

My NNTP WG Notes
$Id: nntp.wg.notes.txt,v 1.1 1996/12/14 21:50:06 sob Exp sob $
By STAN BARBER

Comments on current practices document:

The next release will add streaming as a current practice and remove the
discussion about headers. Chris Lewis with author a draft on RFC 1036 issues.

There are a few problems with the current draft.

1. There was a feeling in some parts of the group that if this document were
to become any kind of RFC (even informational), it would encourage implementors
to implement all the commands described in it. In the next release, a very
explicit statement discouraging this will be made.

2. There is no clarification of RFC977 concerning article numbers. There are
some implementations that will backfill missing article numbers. The group 
seems to think that article numbers should never backfill unless there is 
some kind of "reset" event. However, XXX from Demon Internet will author some
proposed text to address backfilling.

3. The return codes for AUTHINFO GENERIC do not agree with the actual
implementations. Chris Lewis will supply correct information so this can be
fixed.

4. There was a question about why the V2 DATE command was included. I said
that I had saw it in at least two release implementations, so this is not
really a problem.

5. There was a suggestion that some attempt be made to describe how many 
implementations any particular extension is in. While I did make some attempt
at this, apparently more is requested. I will see what I can do.

I said that I expected to release the next version of the document in the
middle of next month.

Comments on the RFC977bis document:

There was alot of confusion about grandfathering the X commands as part of the
new standard. There was considerable sentiment on both sides of this issue. I
believed we had decided that it was perfectly ok for implementors to include
them in their implementations, but that putting them the standards document as
part of the standard was not a good idea.

It was decided by the group to rename the PAT command to HDR and extend the
functionality such that if no pattern is supplyed then an asterisk would be
assumed. 

It was decided to remove AUTHINFO GENERIC and replace it with AUTHINFO 
USER/PASS. However, this was also a suggetion to combined the two and that
AUTHINFO USER/PASS would imply AUTHINFO GENERIC SIMPLE. We decided to 
discuss this further on the mailing list.

There was a suggestion to delete the specific reference to TCP except to
put the statement in the document that over TCP, port 119 would be the
standard port used.

There was considerable discussion about the use of optional commands in the
standards document. There seemed to be consensus that all commands in the new
standards document should be implemented in any implementation that would
consider itself fully compliant.

Someone commented that the initial description of the protocol (during
the description of the various steps) was contradictory. I acknowledged that
and will fix that in the next version.

I do expect to release a new version of the document about four weeks after
the release of the practices document.