xml-dev - Re: "RDF + Topic Maps" = The Future

[Jonathan Borden <jborden@mediaone.net>]

Martin Bryan wrote:

> ...An RDF statement
> normally assigns a set of characteristics to a single resource, either by
> being embedded within that resource or be referencing it.

    Remember that raw RDF is a generic and low level mechanism for
representing statement triples (predicate,subject,object) where the
predicate and subject are resources and the object is a resource or literal.
But realize that RDF defines resource containers (e.g. rdf:Bag). Since
resource containers are themselves referenced as resources it is actually
quite straightforward to attach a property, as you term 'characteristic' to
a collection.

>Topic Maps state
> which sets of resources share a single characteristic. They also has the
> advantage of being able to characterize characteristics (using scopes and
> associations).

    Again the RDF concept which is equivalent to a 'set of resources' is a
collection. Think of RDF as an assembly language for higher level constructs
such as sets of associations or as a platform on which to layer semantics.
An example is DAML-ONT http://www.daml.org/2000/10/daml-ont.html

    Associations in RDF are simply triples. 'Scopes' probably map onto RDF
collections.

>
> > It's downright exciting!
>
> Both RDF and Topic Maps have the same weakness: They are only as good as
the
> semantics they are based on. Neither provides a standardized mechanism for
> recording the meaning of the characteristic
>

    Such a mechanism for recording the meaning of characteristics is termed
an "ontology". This is one of the primary benefits of RDF, RDF Schemas being
fundamentally a language for creating ontologies. DAML-ONT is an extension
of RDFS. A list of such ontologies is available at:
http://www.daml.org/ontologies/ontologies.html

Jonathan Borden
The Open Healthcare Group
http://www.openhealth.org