The current DwC Terms version (at http://rs.tdwg.org/dwc/rdf/dwcterms.rdf) defines the domain for (at least some) object properties.
Unless I am missing something this means that when I use DwC Terms for properties I implicitly assert the class of the subject, which means that either I can use DwC terms only for subjects of the type they are intended for (and why would one want to limit DwC's use in this way?), or reasoning based on RDF or OWL extractions of leads to problems.
For example, in the Phenoscape project [1] we would like to link characters, states, or OTUs to the specimens based on which a systematist defined a (or all) character(s) or state(s). The specimens would be described in a NeXML document [2] with embedded RDFa annotation by institutionCode, collectionCode, and catalogNumber. The RDF extracted from that when run in a reasoner (which we will do) would implicitly assert that the specimen is a DwCTerms:Occurrence. Right now that probably doesn't hurt much in this case, but it might in the future, and I'm not sure what is gained from forcing those implicit assertions in the vocabulary.
Should I post this as an issue to the tracker on Google Code, too?
-hilmar
[1] http://phenoscape.org [2] http://nexml.org