Very timely after Erik Wilde’s talk about conceptual models for XML, Elias Torres cites from Nature Biotech Journal:
The [above] problem originates from the limited expressiveness of the XML language. This claim may appear to contradict the often proclaimed ’self-descriptive’ nature of XML. But XML, designed as a language for messaging encoding, is only self-descriptive about the following structural relationships: containment, adjacency, co-occurrence, attribute and opaque reference. All these relationships “are indeed useful for serialization, but are not optimal for modeling objects of a problem domain.
This is the best articulation of the “RDF vs. XML” argument I’ve seen so far.