W3C home > Mailing lists > Public > www-tag@w3.org > November 2008

Re: HTTP Link Header

From: Jonathan Rees <jar@creativecommons.org>
Date: Fri, 14 Nov 2008 08:30:23 -0500
Message-ID: <760bcb2a0811140530v49cbc86ere86fd697c175da59@mail.gmail.com>
To: "Lisa Dusseault" <lisa@osafoundation.org>
Cc: "Phil Archer" <parcher@icra.org>, "www-tag@w3.org WG" <www-tag@w3.org>

Footnotes / corrections...

On Thu, Nov 13, 2008 at 2:22 PM, Jonathan Rees <jar@creativecommons.org> wrote:
>  RFC2616 seems to imply that a
> 200 response provides an entity that either *is* the resource itself or is
> something that can be substituted for it (a "representation" or "variant" of
> it).

I phrased it this way to provide lots of wiggle room, but went too
far. RFC 2616 is usually careful to distinguish the entity from the
resource, saying "an entity corresponding to the requested resource"
and so on. (The W3C Web Architecture recommendation also makes this
distinction, with slightly different terminology.)

> preferably by using 303 or a 307 to a different server that in turn delivers a 303.

That would be a 303 to a different *resource* (a document that tells
you about the relation), or a 307 (or 301 or 302) giving a URI to
which the same considerations would apply recursively.

The argument against 200 for URIs naming relations doesn't depend on
the details ("document" vs. "information resource" or "entity" vs.
"representation"), it just depends on the idea that there are no
entities "corresponding to" or "representing" things like relations.
The www-tag archives are replete with discussion of this topic (and
everyone's sick of it and will hate me for raising it again), which is
why my aim was only to summarize, not defend.

(Thanks Stuart.)

- Jonathan
Received on Friday, 14 November 2008 13:30:59 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:48:08 GMT