On Fri, 2003-09-12 at 12:26, Roland Schwaenzl wrote:
> \Mikael wrote:
> > The DC metamodel is based on a resource-property-value model, much like
> > RDF.
> > LOM is based on a containment model (elements contain elements that
> > contain elements, etc.). There is no explicit resource-property-value
> > semantics in this model. This makes the LOM metamodel very close to the
> > XML model.
> I'm not quite sure, whether i understand this correctly.
> Do you mean non-global elements/attributes are used and no URIref
> binding for properties/classes exists?
Yes, there are non-global elements, and no distinction is made between
which are classes and which are properties. Also, no URIref binding is
given. The URIref bindings are specific to the respective bindings. The
XML binding uses a different namespace than the RDF binding.
The RDF binding *does* specify which URIs are classes and which are
properties, of course, but without guidance from LOM itself.
> > But mapping LOM to RDF, or trying to combine LOM and DC elements by
> > aligning the LOM elements with the DC metamodel, mut be done on an
> > element-by-element basis, as the metamodels do not match.
> QName/URIref problem?
No, it's more of a problem of finding out which LOM "elements" are to be
interpreted as properties, and which are values, and so on.
> > I could give many examples from the LOM-RDF binding, but I'll stop here