RE: Metadata schemas

Kevin,

In loking at the relationships for keywording, we might find something
useful from:

http://www.w3c.rl.ac.uk/SWAD/thesaurus.html

They have mapped the approaches in the thesaurus community into RDF schema
and are moving on to look at the use of OWL.

	Andy

-------- Original Message --------
> From: Kevin Smathers <mailto:kevin.smathers@hp.com>
> Date: 20 October 2003 18:07
> 
> Hi all,
> 
> To help my own understanding of the relevant schemas, I've mapped out
> the IMS and OCW schemas as UML, which I've attached to this e-mail.  The
> IMS diagram is taken from the RDF mapping of the IMS schema, which is
> basically LOM translated into RDF with some funky bits (I'm not sure I
> like the author's tendency to use namespaces as first-class bits of the
> schema definition).  The OCW schema is just a straight forward mapping
> of the XML metadata in our repository.
> 
> Missing from the IMS schema is any reference to keywords.  The RDF
> mapping document mentions keywords but simply suggests that the
> dublin-core subject property be used to represent keywords.  This isn't
> mentioned in the schema since the schema uses namespaces to represent
> the relevant class and dc:subject isn't within the namespace.  The same
> goes for description and for several other properties unfortunately.
> 
> In thinking over keywords and thus vocabularies I thought a more
> expressive form of subject was needed to represent the navigation
> relationship between keywords since many vocabularies are structured
> into hierarchies of related terms.  The third attachment is a tentative
> model of a Keyword class that would model hierarchical relationships
> among keywords to support vocabularies like the Getty AAT.
> 
> (If anyone wants the source files, they are in CVS under
> corpus/ims/OCW.  The UML diagrams were created using the open source
> 'dia' tool.)
> 
> Cheers,
> -kls