Paul Downey is in fine fettle:
It seems that folks like Yahoo! and Twittervision offer different representations of the same resource using different URIs .xml, .json, .php, .yaml, etc. Iâd say thatâs arisen because users have little control over content-negotiation in their browser. Isnât having the same resource identified by lots of different URIs an issue for the Semantic Web?
Paul quotes the use of owl:sameAs for expressing resource
identity, but suggests the
Content-Location
header as an alternative which doesn't require grokking of
out-of-band
metacrap semantics.
I reckon he has a good point, resource/representation equivalence is something which could/may be expressed in the protocol, potentially available for a broader set of tools. A closely related question was raised on the semweb list recently : IRI meets RDF meets HTTP redirect. There are a few different aspects contained in Jeremy's scenario, some of which might be helped by xml:base, it's quite a convoluted thread, but Sandro returns to the owl:sameAs point suggesting 301s license that identity.
It seems there is a general problem here, there doesn't appear to be a definitive mapping between declarations in HTTP and RDF/OWL (maybe once I've settled in at Talis I can talk iand into me working on a Note about it...). Nearly forgot - there is the HTTP-in-RDF vocabulary, but that doesn't attempt to do more than describing the field values.
There is pretty entrenched skepticism around about certain bits of WebArch on the web at large, at least in part caused by the implementation hassles with conneg Paul mentions (e.g. David Megginson: "Screw Plato! Viva materialism! Go ahead and put â.xmlâ at the end of your URLs."). Getting the identity semantics sorted might at least allow consistent workarounds such widespread (but questionable) practices, a kind of decriminalisation of something that laws could never prevent.
( re. OpenID, it is a real gift for RDF tools)
btw, Paul has put a good photo of the Linked Data session at WWW2007 up on Flickr, if you were there you might like to identify yourself...
@en