RE: NEW ISSUE: cacheability of status 303

> From: Roy T. Fielding [mailto:fielding@gbiv.com]
>
> On Oct 16, 2007, at 6:23 AM, Julian Reschke wrote:
> > Booth, David (HP Software - Boston) wrote:
[ . . . ]
> >> BTW, I do notice one other thing.  I suggest changing the
> >> following sentence:
> >>     A 303 response to a GET request indicates that the requested
> >>     resource does not have a representation of its own that can be
> >>     transferred by the server over HTTP.
> >> to:
> >>     A 303 response to a GET request indicates that the requested
> >>     resource does not have a representation of its own, available
> >>     from the request URI, that can be transferred by the server
> >>     over HTTP.
> >> The reason is that if the same resource were requested via a
> >> different URI, it might indeed provide a representation of its own
> >> (if it were an information resource).
>
> In which case it would be redirected via a 301, 302, or 307.
> 303 only redirects to different resources, which means the requested
> resource for the 303 response is different from the target resource,
> even if that difference can't be measured in bits.  Even if they
> aren't, in fact, different, the client is being told by the server
> that they should be interpreted as different, and that makes it fact
> as far as HTTP's interface is concerned.

That isn't the case I meant.  Sorry I was unclear.  I meant a case where Ua, Ub and Uc are all different URIs and:

 - Ua denotes an information resource R;
 - Ub denotes the *same* information resource R; and
 - Uc denotes some other information resource.

When Ua is dereferenced, a 303-redirect to Uc is returned, but when Ub is dereferenced a 200 Okay is returned with a huge amount of data.  And when Uc is dereferenced, a 200 Okay is returned with content describing R and how it can be accessed via Ub.

>
> There is no information resource in HTTP, for the same reason that
> there is no spoon in the Matrix.

Yes, I realize that.  I am using one in this example to highlight how the HTTP protocol specification interacts with the WebArch.


David Booth, Ph.D.
HP Software
+1 617 629 8881 office  |  dbooth@hp.com
http://www.hp.com/go/software

Opinions expressed herein are those of the author and do not represent the official views of HP unless explicitly stated otherwise.

Received on Wednesday, 17 October 2007 12:43:07 UTC