> This requires ontologies to be parsed to N3 and back every time EYE is called, correct?
Yes. For the first part (producing N3), this is not a problem since NTriple is a subset of N3. Thus any NTriple file is a valid N3 file. So NTriple files produced by Owlready should be OK pour EYE.
The second part (N3 parser) requires to write a dedicated parser. Implementing N3 in Owlready would be a nice addition, but I'm very short on time currently... Any contributions are welcomed, though ;)
An alternative is transform N3 files in NTriple or RDF/XML. I think this can be done in Java with OWL API.
MCF, LIMICS, Université Paris 13