Thanks to Paul, Hilmar and Bob, the advice is useful to us.
- When documenting legacy xml ontologies, the xml variant requires
documenting a difference between element and attribute children.
What's an XML ontology? Isn't that an oxymoron?
I personally think an xml schema represents an ontology in a generic sense. It has only a subset of functionalities, just like a thesaurus from the 1960 is a useful subset of the functionality of current ontologies. But Owl has only a subset of Owl2, and I suspect this will be a subset of the future.
But our task is actually: just read "documenting a publishing dtd or xml on a semantic wiki". The goal is to keep as much orthogonal as possible, whether an RDF vocab or an xml schema is documented. That does not mean that everything has to match.
To my limited understanding, both attributes and child elements in xml schema/DTDs are a class-membership relation. Is a membership, say in any kind of UML model a part-of relation? I suspect so. Incorrect?
And then we want to express something about the kind of relation, which in RDF requires a reification, which in semantic mediawiki we don't (at least not easily). So I wonder what to do, to keep it simple. The easy way is to define an xml schema class with our own special properties, but we would like to re-use existing vocabularies where possible.
Gregor