> From [log in to unmask] Fri Sep 27 12:29 MET 2002 > RDF-XML: <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" > xmlns:dc="http://purl.org/dc/elements/1.1/"> <rdf:Description > about=""> <dc:publisher> UKOLN - University of Bath </dc:publisher> > </rdf:Description> </rdf:RDF> > DC.Publisher: UKOLN - University of Bath > MIME-Version: 1.0 > Date: Fri, 27 Sep 2002 11:19:55 +0100 > From: Andy Powell <[log in to unmask]> > Subject: Re: multiply affiliated refinements > To: [log in to unmask] > > On Fri, 27 Sep 2002, Roland Schwaenzl wrote: > > > Dear Andy, > > > > i don't think your idea works: > > > > You propose > > > > dc:creator --subPropertyOf--> dc:contributor. > > > > That would imply any refinement of dc:creator is automatically > > a refinement of contributor. > > > > That is you create the problem you want to have forbidden to happen! > > No. I said that each DCMI element refinement should have only one parent > DCMI element or element refinement. So, supposing someone proposed a > dcterms:creatorIllustrator property as a refinement of dc:creator. Then > dc:creator would be its DCMI parent property, dc:contributor would be its > grand-parent property. Dumb-down would work fine - everyone could go home > happy and sleep soundly in their beds :-) > Aaah! You want something like nearest relative. rs