Dear colleagues,
I have just spent indefinite time staring at a single sentence from the "National Information Exchange Model Naming and Design Rules", section 5.4:
"The properties of an object may be spread across several XML elements that have the same identifier." [1]
The significance of this sentence for the potential of XML technology is hard to assess. XML technology is based on the XDM model, which is centered around the concept of node identity. We might add to the XDM a second concept of identity - semantic identity - which is shared by all nodes describing the same resource (in the RDF sense), in other words, all nodes which are part of the same *object*. (The adherence of an XDM node to a semantic object is simply signalled by a special, URI-valued attribute, like NIEM's structures:uri). Note that the nodes of an object can be distributed over distinct documents. Consider enhanced navigation, for instance, a new axis object:: which maps the context node to all nodes sharing its semantic identity. Michael Kay has more than once complained about the arbitrariness with which information is encapsulated by particular documents whose identity should not be our concern (at least, if I understood some remarks correctly). Please note how the introduction of semantic identity introduces a novel, semantic coherence of information which is independent of resource borders.
Hans-Jürgen
