- From: Roy Fielding <fielding@beach.w3.org>
- Date: Wed, 09 Aug 1995 15:04:20 -0400
- To: Dave Kristol <dmk@allegra.att.com>
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
>Maybe I'm incredibly dense (I see those heads nodding :-), but I found
>Sect. 8.28 a bit confusing.
No, not dense, that was the last section added in response to Jim's
earlier comments that the URI header was insufficient for caching.
This is the first time the text has been reviewed by anybody other than me.
>What would have helped me is for there to
>be an explicit statement that particular vary-dimension's correspond
>exactly to specific HTTP headers, e.g.,
> type Accept
> charset Accept-Charset
> language Accept-Language
> encoding Accept-Encoding
> user-agent User-Agent
> version ??
>The examples imply the correspondence but there's no explicit statement
>of correlation.
Yep, will do.
>What, exactly, is "version"? Where is the version specified in the URI?
Ooops, that's a leftover -- version is not necessary and will be deleted.
Versions are specified by different URIs.
>How does a caching intermediary handle a vary-dimension of <token>?
>That is, if you can't specify the dimensions in advance, how can the
>intermediary figure out what the semantics for such an unspecified
>dimension are? (If <token> is just meant as a place-holder for future
>expansion, let's say so.)
Actually, I was going to remove token altogether, since these things
can't be extended without changing the protocol version.
....Roy T. Fielding Department of ICS, University of California, Irvine USA
Visiting Scholar, MIT/LCS + World-Wide Web Consortium
(fielding@w3.org) (fielding@ics.uci.edu)
Received on Wednesday, 9 August 1995 13:48:23 UTC