I posted a more extensive discussion of the issues you raise.
I hope that we might take this discussion off the Protege - owl forum, but
we are interested in continuing to talk about and reveal the issues related
to translation of other paradigms, such as the OASIS SOA-IM (service
oriented architecture information model) to OWL. So perhaps some discusison
here will be tolerated.
While we realize that Protege is a research environment, the current
critical market need is for interoperability between the W3C based
(lay-cake) logic based ontology, and other representations of human or
computer transaction spaces - such as IEEE and OASIS information models and
metadata registries, such as the ISO/IEC 11170 - which has been recently
discussed on the Protege discussion forum.
The effort my group is making exposes some difficulties in simply taking a
simple RDF file, produced using the new commercial Altova RDF editor, and
importing this to Protege as pure RDF. It is likely that the difficulty
stems from our misunderstanding of how namespaces are used, but that is
conjectural on my part.
Andrea has made several very helpful communications regarding the nature of
RDF and RDF schema (RDFS) and the notion of a "language". I am posting some
of his discussion along with other's contributions into the BCNGroup Galss
Bead Games - for the record.
The import-export issue arises in other context, such as the import of the
BioPAX ontology. This import can be done but one has to understand a
complicated process having many decisions, where each decision requires the
"researcher's" understanding of the state of the current version. It is a
moving target, which I am sure is well known by the Stanford team members.
I do not speak for the BioPAX working group. But I know that they have
chosen to use Protege-OWL rather than Protege-frames for various reasons, in
spite the traditional choice of Protege-frames for bio-medical ontology.
Part of the reason is that BioPAX is designed to be consistent with the
notions that Alan Rector uses in "Defaults, Context, and Knowledge
Alternatives for OWL-Indexed Knowledge bases" ...
".. despite the benefits of logic-based formalisms such as OWL, the
inability to express defaults and exceptions is a serious limitations"
One can down load this paper and read the full paragraph, and text.
Many individuals see the limitations, not only as discussed by Alan Rector,
but also the deeper root causes of these limitations as expressed in the
philosophy of mathematics.
In spite of this, we see the possibility that OWL constructions are "almost"
ready to instrument a "Semantic Web" so that (quoting Rector) "the use of
the logic based ontology as an index to contingent information ..." is
available within various fields including medicine.