 |
 |
On Fri, 9 Aug 2002, Chris Croome wrote: > I have been thinking about the usage of conformsTo I guess that this how > it would be used for RDF: > > <dcterms:conformsTo> > <rdf:Description rdf:about="http://www.w3.org/TR/xhtml1/"> > <dc:title>XHTML 1.0 The Extensible HyperText Markup Language</dc:title> > </rdf:Description> > </dcterms:conformsTo> > > And this for XHTML: > > <meta > name="DC.Relation.conformsTo" > scheme="URI" > content="http://www.w3.org/TR/xhtml1/"/> Sounds reasonable... and this for XML <dcterms:conformsTo xsi:type="dcterms:URI"> http://www.w3.org/TR/xhtml1/ </dcterms:conformsTo> > What I'm not so sure about it what URI to use? > > When producing metadata for a XHTML document should the URI be the > specification, as above, or the namespace or the DTD? Namespace feels wrong to me - since for some/many 'specifications' there will be multiple namespaces. Furthermore, not all 'specifications' will be XML or SGML based... e.g. Web Content Accessibility Guidelines 1.0... so linking to the DTD or XML schema won't be possible. Therefore, in general I'd use the 'preferred' URI for the human-readable form of the specification that is being conformed to. E.g. in the case of WAI, I'd use one of http://www.w3.org/WAI/WCAG1A-Conformance http://www.w3.org/WAI/WCAG1AA-Conformance http://www.w3.org/WAI/WCAG1AAA-Conformance (based on the guidance at http://www.w3.org/WAI/WCAG1-Conformance). For your XHTML example, I'd use http://www.w3.org/TR/xhtml1/ as you have. In some cases, there won't be an appropriate URI, in which case I'd use the 'title' of the specification. E.g a resource that conforms to GCE A level (an educational level in the UK) one might simply have to use GCE A Level In general, I see the values of DC elements (and element refinements) as being human-readable. A URI for a spec is a useful identifier... clearly a URI is machine-readable, but in general, I'm not convinced that DC consuming applications will want (or expect) to be able to retieve what is at the URI and do anything useful with it (other than display it to the end user, or allow the end user to 'click' on the link). That said, given any two descriptions, it would be nice to be able to tell if the resources both conform to the same spec or not - therefore, having some consistency in how we 'name' specifications is very valuable. Using the URI for the human-readable spec feels like the 'best' approach to me. > The 'Nature of HTML' section of this page: http://www.rddl.org/natures/ > discussed this. > > Thoughts? > > Chris > > -- > Chris Croome <[log in to unmask]> > web design http://www.webarchitects.co.uk/ > web content management http://mkdoc.com/ > everything else http://chris.croome.net/ > Andy -- Distributed Systems, UKOLN, University of Bath, Bath, BA2 7AY, UK http://www.ukoln.ac.uk/ukoln/staff/a.powell +44 1225 383933 Resource Discovery Network http://www.rdn.ac.uk/