Comments on: 303 Asymmetry

By: Ed Davies

Ed Davies — Thu, 06 Dec 2007 10:32:52 +0000

P.S:

In a message on the TAG list you wrote:

“Yes. We better hurry up and get the link header through the ietf – there
are more and more uses for it every day ”

I fully support that.

I rather like TimBL’s N3 header for power but any sensible design’s better than ignoring the matter and having a heap of kludges with content negotiation and so on crop up.

By: Tore Eriksson

Tore Eriksson — Thu, 06 Dec 2007 00:09:42 +0000

(Reposting since the previous post had the markup removed. Please delete it if possible.)

My non-kosher solution:

GET /R HTTP/1.1
Host: example.org
Accept: application/rdf+xml

HTTP/1.1 200 OK
Content-Location: http://example.com/RDESC
Content-Type: application/rdf+xml

…

http://example.org/RDESC) as subject –>
The RDF document.
Ian Davis

—

I know this is NOT what the TAG recommends, but I still can’t see any semantic problems with this approach. Maybe I think too much in REST terms, but I do not have any problems with serving representations where the information content differs, as the mime-type says what sub-part of information the representation contains.

Tore

By: Richard Cyganiak

Richard Cyganiak — Wed, 05 Dec 2007 13:32:48 +0000

Why not embed the description of RDESC into RDESC? Or publish it as a separate document and link to it from RDESC? After all, RDESC is a document whose content we can GET, unlike R.

(I realize that you are looking for solutions at the level of web architecture. I just want to point out that there are simple yet limited fixes that work today, and perhaps would work well in the Talis scenario you describe above.)

By: Ed Davies

Ed Davies — Wed, 05 Dec 2007 12:34:31 +0000

Ian,

There was an earlier discussion on the TAG list about this. My question:

http://lists.w3.org/Archives/Public/www-tag/2007Aug/0075.html

and TimBL’s reply:

http://lists.w3.org/Archives/Public/www-tag/2007Aug/0089.html

It looks like everybody comes up with a slightly different idea for the details of the header(s) but we’re all thinking basically the same thing.

As to whether they’d have come up with 303 if some such header had existed – I don’t know. But use of 303 and such headers are not exclusive, though combination would perhaps be a bit more complicated than absolutely necessary.

Ed.

By: Simon Reinhardt

Simon Reinhardt — Wed, 05 Dec 2007 07:01:11 +0000

>> I seem to recall something similar to this being proposed a few years back but my Googling doesnâ€™t turn anything up.

Ian, I think what you were looking for is http://esw.w3.org/topic/LinkHeader .