Is there a way of – inside an RDF document – specifying a physical URL for URI references? That is, something that does the same job as xsi:schemaLocation in XML schema documents?
I understand that the “correct” way to do it is with the URI doing a 303 redirect with content negotiation, but if setting up redirects is a hassle, can my RDF document itself explain that
http://foo.bar.com/voc/SpecialSettings
is to be fetched from the physical location
http://server1.bar.com/prod/app3/SpecialSettings.owl
?
Update: we resolved this by explaining to the server that it needed to send redirects. See http://biodiversity.org.au/voc/ibis/IBIS
This is not possible with standard RDF. AFAIR, this additional complexity was deliberately avoided.
You could cheat and say in an ontology that http://foo.bar.com/voc/SpecialSettings is owl:sameAs http://server1.bar.com/prod/app3/SpecialSettings.owl. Then a program that understands owl:sameAs might try both URLs and combine the information it gets from both, which in that case would be nothing from the first one and everything from the second one. But not many programs and libraries will do this automatically. It’s also fishy from an ontological perspective. You, as the data provider, should provide one canonical URL for your resources.
For this kind of question, my first reference is How to Publish Linked Data on the Web.