Re: HTTP WG status, scheduling

Koen Holtman wrote:
> 
> Larry Masinter:
> >
> >Content negotiation:
> >   There are several new drafts, and we have yet to absorb them.
> >   there's been some discussion.
...

> Ehm, the hit metering draft does not belong here.

Right.

> >        (draft-mutz-http-attributes-02?)
> 
> I have different comments on the timing of the different drafts:
> 
> 1) draft-ietf-http-negotiation-00.txt
> 
> As per the decisions at the December IETF, this draft contains the
> non-controversial stuff which can be moved forward quickly, so I suggest we
> move it forward quickly.  There are no big changes with respect to the
> previous version; everybody who wants to absorb the changes should be able
> to do so on short notice.

One might hope so. But on the other hand, people are busy.

> I'd like a last call for this draft as soon as possible, say the end of this
> month.  I feel that it is high time to get the Alternates header format onto
> standards track, so that servers can start using it.  The lack of a standard
> Alternates header format has blocked progress in client-side negotiation for
> too long already.

I'm hoping everyone will comment on the entire schedule, but I agree,
we can do these in order.

> 2) draft-ietf-http-rvsa-v10-00.txt
> 
> This draft contains the part which was deemed controversial at the December
> IETF.  The algorithm should now be less controversial because it is made an
> optional adjunct to the main draft: other algorithms can be defined if
> wanted.  I have never seen any convincing example (here or on the i18n list)
> of a case where this algorithm would break.  I propose to do a last call on
> this after 1) has gone through.  If the draft remains controversial, we may
> consider making it an experimental RFC.


I'd like feedback from HTTP client/server/proxy implementors on this.

> 3) draft-ietf-http-feature-reg-00.txt
>    (draft-mutz-http-attributes-02?)
> 
> Feature tag registration and the core feature set are still very much works
> in progress, unlike 1) and 2) above.  Setting up a feature negotiation
> subgroup to move forward this work might be a good idea.  I think the May
> 1997 timing for this work is reasonable.

I'd like to hear from others on this, too. I think that "device
characteristics"
are important protocol elements not just for HTTP but for other
protocols
that want to do content negotiation. IPP and Internet Fax are both
discussing
protocols that need content and feature negotiation, even when not cast
as
part of HTTP.

Larry

Received on Sunday, 16 February 1997 15:11:17 UTC