I'm a bit disappointed with owl2.  Just stepping back, considering the domain from the perspective of an internet technologist unfamiliar with semantic web technologies, it's probably confusing that the language used to order, model, or govern the semantic web (which is to be composed of RDF in various manifestations, RDFa for example) won't use RDF itself, and requires tooling to translate between itself and the lowest common denominator of the sem web.

I'm aware that the working group is composed of experts who know much more about this stuff than I do, but I fear that OWL2 will dampen the growth of the practical semantic web unless it's extremely complimentary and supportive of RDF.