Hi Matthew - I think you have a small understanding. OWL 2 has two
interpretations : Direct (i.e. DL) and RDF-based (more like FOL) and
so you don't have to stick with the DL limitations for 15926 as OWL.
That's why there will likely be at least two representations in the
PCA project ... one or more wearing the DL straightjacket to support
reasoning, and one unencumbered by DL to support data integration.
MW: I understand that OWL 2 still has the same single
level of classification that UML and EXPRESS have in
principle. So you either have to do what we did, as you say,
define your own language using another – and loose native
support, or make some compromise. Neither of which is ideal,
hence not really satisfactory. That does not mean such
attempts are without value.