W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > September to December 1995

Re: On the connection between content negotiation and caching

From: Daniel DuBois <ddubois@rafiki.spyglass.com>
Date: Thu, 28 Dec 1995 14:24:13 -0600
Message-Id: <9512282024.AA07168@rafiki.spyglass.com>
To: "David W. Morris" <dwm@shell.portal.com>, http working group <http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com>
>> checking the content negotiation algorithm.  You can't send back text/html
>> to someone who sends "Accept: text/html;q=0.0".

>The current draft 1.1 does not seem to support this position.  By my 
>my reading it is correct to send whatever the server has if it 
>have only one choice.  There is no special significance to q=0.0 that I
>could find. Just really undesired by the requestor.

Well, I get this position from page 73, "If no variants remain with a value
of Q greater than 0 the server should respond with a 406 None Acceptable."
Additionally to support my interpretation there are references to empty
Accept-Charset and Accept-Encoding headers allowing user agents to refuse
resources (no mention of a multiple-variant requirement), so there is some
implication that those headers must be looked at even for non-varying resources.

But given that Content Negotiation is labeled an optional feature, that some
of the statements int the spec could be interepreted the way you did, and
that some people may not feel negotiation applies to non-varying resources,
I realize the intention of the spec might not be as I had perceived.  If
that's true, I think it should be made more explicit one way or the other.
-----
Dan DuBois, Software Animal             http://www.spyglass.com/~ddubois/
		I absolutely do not speak for Spyglass.
Received on Thursday, 28 December 1995 12:29:30 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:38 EDT