notes from previous Editing Group meeting: goals, docking

In last week's telephone conference, we discussed the
editing group goals:

To move to "Draft Standard", we must assert that there are
two interoperable independently-derived implementations of
each feature. (The exact criteria are in the 'Internet Standards
Process' RFC.)

While some would like the HTTP/1.1 spec to be shorter, we're
going to emphasize clarity over brevity.

We plan to create an index/table of features and requirements
in the same manner as RFC 1122,1123, in order to make clearer what
is and isn't required of clients, origin servers, and proxies,
repectively. There will be an example of this soon.

We'll review the wording of MUST/SHOULD etc. against the
new BCP that describes the use of those words, and bring
the spec into alignment as needed.

We discussed our proposals for "docking": whether a work item
is or is not part of the HTTP/1.1 specification. This doesn't
necessarily mean 'in the same document' but 'as part of
the same suite of documents'.

Our proposals: 
- State management won't dock
  It will be on its own track, and might even have a
  separate working group 
- Security will not dock
  This wasn't really seriously considered, but we confirmed
  that SHTTP and HTTPS are independent specifications
  and/or standards. 
- Digest authentication will "dock"
  We might separate out 'authentication' as a separate RFC,
  and 'link' it, but we're considering it one spec even
  if released as a suite. 
- Whether PEP will dock depends on progress on PEP. 
- Hit Metering will not dock 
- Content negotiation
  We concluded (after a lengthy discussion) that we felt
  the content negotiation work needed a separate focussed
  effort, with a 'requirements document' as well as a
  specification. 

Some of these are admittedly controversial, but in
perspective, this list is consistent with getting
a stable HTTP/1.1 Draft Standard soon.

Larry
--
http://www.parc.xerox.com/masinter

Received on Friday, 23 May 1997 18:26:40 UTC