Minutes for CONNEG for the 47th IETF.
Submitted by Ted Hardie, Chair
Reported by Geoff Horne

The group met Monday, March 27, 2000 from 7:30 to 9:30.  This meeting
should be the last meeting before the group closes down.  CONNEG has
only one milestone item to complete (short form representation of
feature predicates); once it has completed that item, it will close.


The group first reviewed related work:

draft-hoffman-char-lang-media-02.txt (last call requested and expired)
draft-itef-fax-content-negotiation-01.txt (in early stages of completion)
draft-itef-rescap-proto-format-00.txt (may carry conneg data, but no requirements)
w3c CCPP (working on RDF-based system for capabiliti negotiation)
WAP UAPROF (input into CCPP and a format for client capability information within WAP)

Graham Klyne then reviewed comments from the IESG on the type feature
and the content-features header proposals.  There were no objections
to the clarifications proposed.  Only one substantive change is
involved: where the type draft currently strongly discourages using
parameters to MIME content-types listed in a type header, they now
MUST NOT be included.  Other changes to the draft included improved
examples, which will be submitted in the revised draft to the IESG

Ned Freed pointed out that it would be useful to have a content type
for media features, and he agree to write the document registering
the type.  He set a goal of March 31st, 2000 for the registration.

The group then reviewed the goals of the short form predicate proposals:

1) Do no harm.  The current system works, but is chattier than it could be;
this work is an optimizaiton and should be considered in that light.

2) Provide a method for using a short form to replace a number of
features which occur in several alternate predicates.

3) Provide a short method for referencing a common feature grouping
where there is a restricted set of such groupings such as occurs in
the fax case.

4) Provide a common reference method for the short forms of feature
predicates and the other feature types or values.  This implies
standardization on URIs, because of the u. features.


The group first discussed the common reference method issue and agreed
that a URN-based reference for the registered features would be
valuable.  The group also agreed that this would be best accomplished
in an IANA-wide registry and the chair agreed to discuss the matter with
Michael Mealing and Leslie Daigle of the URN working group. [This
discussion occurred later during the week and Michael Mealing agreed to
write the documents necessary for creating an IANA URN name space).

The group then moved into a discussion of how to handle the short form
issue for feature predicates.  A lively debate ensued, with three
examples from Graham Klyne providing the focus for several use cases.
There was considerable confusion in the group present on the
distinction between two different cases in which URLs figured as part
of the auxillary predicate.  During this discussion the document
authors disagreed on the need for dereferencing in one of the cases,
and one of the document authors withdrew support for cases in which
dereferencing is necessary because the operational characteristics of
that case were likely to be sub-optimal.

After a discussion period did not produce agreement among the group,
the chair asked the document authors to take the in-line method upon
which the group agreeed and seperate it into a different document.
That document will be advanced to the IESG as a candidate for proposed
standard.  The document authors will revise the other two methods
within a new document.  The current plan would be to submit that work
as an individual submission either at a later time or for experimental
status now.  Graham agreed to have the seperated draft to the working
group for review by April 10, 2000 with a goal of working group review
complete by April 30, 2000.