The editorial group working on the new HTTP/1.1 draft has agreement on
the following strategy to deal with content negotiation.

If anyone strongly disagrees with this strategy, now is the time to
speak up.

I will soon post draft text that reflects this strategy.

The main goals of the strategy are 1) to avoid the risk that
disagreement about content negotiation would delay consensus on the 1.1
draft until after May 1, while 2) not blocking a smooth introduction
of content negotiation after May 1.

The plans are as follows:

- as announced earlier in a message by Jim Gettys, (Proposed structure
of HTTP 1.1 document(s), Mon, 18 Mar 96 15:03:18 -0500), the content
negotiation text in draft-holtman-http-content-negotiation-00.txt will
not be merged with the main HTTP/1.1 document.  An updated version of
draft-holtman-http-content-negotiation-00.txt will appear in April,
and it is hoped that consensus on it can be reached in May.

- Section 12 of the old 1.1 draft (which gave an incomplete definition
of a content negotiation mechanism) will not be present in the new 1.1

- There will be a Vary header in the 1.1 draft, so that HTTP/1.1 can
support opaque negotiation on language in a reasonably efficient

- There will be Accept* headers in the 1.1 draft.  With respect to the
Accept* headers previous draft, numerous small improvements are made;
these improvements reflect the consensus of the content negotiation

- There will be `hooks' in the 1.1 draft to ensure that all HTTP/1.1
caches will be compatible, though not in an optimally efficient way,
with a transparent content negotiation mechanisms like the mechanism
defined in draft-holtman.  Thus, transparent content negotiation
(which is what Section 12 of the old 1.1 draft covered incompletely)
won't have to wait for HTTP/1.2 if HTTP/1.2 turns out to take too
long, it can be done on top of HTTP/1.1. 

- The `hooks' for transparent content negotiation consist mainly of an
Alternates header definition which defines the Alternates header as
synonymous with a certain Vary header.  Also, some language in the
draft will announce that a negotiation mechanism using Alternates is

Note for the content negotiation subgroup: I previously announced to
you that I would try to keep a complete Vary header definition out of
the main 1.1 draft, because of concerns that it being in the main
draft could prevent the Alternates header from becoming more popular
than Vary because Vary `got there first', even though Alternates is
much better than Vary in 95% of the cases.  I failed: the editorial
group voted on it, and it turned out that I was the only one one who
thought that putting Vary in the negotiation draft would be better.


Received on Monday, 1 April 1996 13:49:45 UTC