Minutes for CONNEG, 44th IETF
Reported by John Dykstra, edited by Ted Hardie

The group discussed the framework and terminology draft and decided
that it would be useful to have it permanently available as an
introduction and terminological reference, especially given the
differences between X.500 definitions and CONNEG usage.  After
author review for currency, the document will be last called with
a proposed status of Informational RFC.

The group then reviewed the current draft on the "type" content feature.
There was no dissent on the need for a feature which expresses MIME
content type, as this allows more comprehensive set construction for
features which are specific to content types.  The examples in the
document need work; Bob Harriet and Larry Masinter agreed to provide
examples based on specific devices.

The group discussed proposals for a "Content-feature" header.  The
discussion revealed that the group did not have enough the appropriate
experience with the MIME specifications to be certain of the examples.
External review of the examples and proposal will be sought by the
chair.

The group then considered the two proposals for aggregating features.
Both proposals allow for dramatically shortened feature expressions
for complex cases; the set of IE4 capabilities was presented as one
such common capability set that would benefit from these proposals.
The URI proposal presented by Bill Newman allows for configuration at
a variety of levels (original equipment manufacturer, OEM, user),
where the hash proposal presented by Graham Klyne works best in
situations where there is a finite number of well-known sets. Both
have valuable points, but the group was unable to come to consensus on
which approach is best.  The group will continue to work toward a
syntax which combines the advantages of both, possibly by using
secondary references and possibly by presenting a hash-based URI.
The group does believe that these differences are resolvable, but
feels more time with the proposals is necessary to reach resolution.