RE: Variant-ID proposal

I agree that Alternates _should_ be completely orthogonal to Vary. For
example, I would add "User-Agent" and "Cookie" as one of the
discriminators in the Alternates header if needed to make them
orthogonal.

I don't know if I buy the implication that Alternates is only for
reactive content negotiation. Once a cache has the list of alternates
for a resource, it should be able to use it in preemptive negotations,
right?

>----------
>From: 	Roy T. Fielding[SMTP:fielding@avron.ICS.UCI.EDU]
>Sent: 	Sunday, April 14, 1996 10:35 PM
>To: 	Koen Holtman
>Cc: 	http-caching@pa.dec.com; jg@w3.org
>Subject: 	Re: Variant-ID proposal 
>
>Argh, no, you don't need any of this stuff.  The Alternates mechanism
>must be completely orthogonal to Vary.  Alternates provides a means for
>reactive negotiation (i.e., retrieval of a better resource after the
>first response).  Vary provides a means of identifying when the server
>has engaged in some form of preemptive negotiation.  It is both
>possible
>and quite reasonable for both to appear in a single response, meaning
>that the server has chosen what looked to be the "best", but is
>also providing Alternates so that the user agent can later choose
>something "better".
>
>Vary never affects Alternates and the variant-id part of an EID is
>only present when receiving a preemptively negotiated response from
>the server.  A server that only supports Alternates will not be
>serving preemptive responses, and thus will not NEED any variant-ids.
>
>........Roy
>

Received on Monday, 15 April 1996 19:00:56 UTC