Minutes, CONNEG Working Group  
Dec 13, 1998
Ted Hardie, Chair
Reported by April Marine

The chair first clarified the focus of the group for first time
attendees, highlighting its mission to create a registry for elements
of exchange and a syntax for senders and recipients to use in
identifying the intersection of capabilities and available features.

Graham Klyne reviewed the current syntax docs.  Recent issues on the
mailing list have focused on q-values and set notation; those and
editorial changes were considered and the room approved of the
proposed changes.  These will be folded into a final draft and that
draft will be last called on the mailing list.

During the course of discussion, two architectural principles for
registered features emerged.  The first principle is that the
semantics of a feature should not depend on the value of any other
feature.  For example, if you have a feature "cloth" which can be
equal to "cotton", "wool", or "blend" and second feature "washtemp"
which can be equal to "hot", "cold", or "warm", the values of "hot"
should not change in a washing context depending on the cloth chosen.
The related architectural principle is that absolutely determined
features make it easier to reuse the same terms independently;
"washtemp", in other words, should be the temperature in degrees
celsius, not tokens like "hot", "warm", and "cold".

The group briefly considered the CCPP crosswalk draft and noted that
the primary difference between the frameworks is that the CCPP draft
presumes an aggregated feature set based on device; the CONNEG
framework presumes that feature sets get built out of individual
feature collections at the time of exchange.  The two groups have
agreed to use the same uri syntax for uri-based features; the CCPP may
eventual use registered feature sets, when they have been fully
specified.

Larry Masinter presented a possible resolution to the color issue in
the media-features draft; he will issue a new draft, which will be
last called on the list.  The group recognizes that this document is
not meant to be comprehensive, but merely a basic set of useful media
features; the color model and other aspects here are, therefore
necessarily limited.

The group heard a presentation from Lloyd McIntyre and Graham Klyne on
the FAX feature document; this is being considered in the FAX working
group and those interested in that arena should follow the document
there.

Some final questions on IANA considerations were discussed and Larry
Masinter identified as an Area Expert for the IANA to consult in the
area of feature collections.

Further work on feature set aggregation is needed; Chris Burke has
tenatively volunteered to edit that draft.  2/28/99 set as a potential
date for that work.