- From: Ian Horrocks <ian.horrocks@comlab.ox.ac.uk>
- Date: Wed, 29 Jul 2009 23:44:58 +0100
- To: Sandro Hawke <sandro@w3.org>
- Cc: Mike Smith <msmith@clarkparsia.com>, Alan Ruttenberg <alanruttenberg@gmail.com>, "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, public-owl-wg@w3.org
On 29 Jul 2009, at 20:28, Sandro Hawke wrote: >> On Wed, Jul 29, 2009 at 14:20, Alan >> Ruttenberg<alanruttenberg@gmail.com> wrot >> e: >> >>> On Wed, Jul 29, 2009 at 2:05 PM, Ian >>> Horrocks<ian.horrocks@comlab.ox.ac.uk> wrote: >>>> At least HermiT and Pellet are able to operate completely >>>> independently -- >>>> they can take an OWL 2 ontology in RDF/XML format and check if >>>> it is >>>> satisfiable. The use of Mike's test harness is merely a >>>> convenience. >>> >>> What are they using for parsers? Given the setup with P4 as example, >>> my expectation would have been that they both use the OWLAPI to the >>> parsing. >> >> Pellet supports parsing OWL 2 RDF/XML from either its OWLAPI or >> Jena interfac >> es. > > Good to hear. When using the Jena interface, you're saying Pellet has > its own complete implementation of the mapping-to-rdf spec? I presume that what you are talking about is the mapping from RDF graphs to the OWL structural specification. It is perhaps worth pointing out that conformant OWL 2 tools are not in any way obliged to implement this mapping -- they are only obliged to take OWL ontologies in RDF/XML syntax as input and correctly compute certain inferences. They could do this by operating directly on the triples, by translating into some other formalism, by magic, or whatever -- it doesn't matter so long as they give the correct answers. The mapping is simply a way to define the relationship between the RDF/XML syntax and OWL structures, which on the one hand is used to define the direct semantics and on the other hand is used to define document conformance (a conformant DL document is one that *could* be correctly parsed using this mapping). Ian > Have you > compared the two parsing systems, side by side, on all inputs, or > anything like that? > > -- Sandro
Received on Wednesday, 29 July 2009 22:45:40 UTC