- From: Neubert Joachim <J.Neubert@zbw.eu>
- Date: Fri, 23 Jan 2009 17:26:06 +0100
- To: "Dan Brickley" <danbri@danbri.org>
- Cc: "Antoine Isaac" <Antoine.Isaac@KB.nl>, "Houghton,Andrew" <houghtoa@oclc.org>, <public-esw-thes@w3.org>
Dan, > With respect, I suspect this mixes up the graph of nodes/relations > encoded in an RDF document with the graph of document/link relations > in the Web. "Follow your nose" in this context w.r.t. RDF and linked > data means, as I understand it, that unknown URIs can be de-referenced > to learn how to interpret them. This is regardless of whether they > appear in statement. May be I'm somewhere blindfolded. What I want to achieve is: Starting from a XHTML+RDFa page describing one concept, navigate to a search result page of items indexed with this concept. And this link or relation should be expressed both in HTML (which is trivial) and in a RDF(a) property (which seems to bee hard). I see this as keeping in parallel the human and the machine visible layer, rather than mixing both layers up. I'm aware that I can introduce the inverse relation (quite easy by use of a "rev" attribute), but that's not what I want to say in the first place, and I'm not sure how current or future applications will handle this. But maybe it is only my personal feeling that it should be possible to express the relation "indexes" or "isSubjectOf" directly. The use case should be valid even then, and I think it could be helpful to add some guidance in the SKOS primer on how to handle it.
Received on Friday, 23 January 2009 16:44:49 UTC