- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Thu, 27 Mar 2003 15:38:45 +0100
- To: <herman.ter.horst@philips.com>, "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
- Cc: <www-webont-wg@w3.org>
> >Second, OWL DL does have a specification that I believe meets > >the requirements. > > This has been extensively discussed. > The normative specification does not provide direct "syntactic > conditions" as mentioned above. > In order to decide whether an RDF graph is OWL DL or OWL Lite > by means of the normative specification, a software system > would need to use the mapping rules backwards, which would > be a very cumbersome process, if at all possible unambiguously. > > While I agree with Herman that some description other than the mapping rules is needed in Proposed Rec, I do not see an in principle problem with going to Last Call with this missing. No one who approves the Last Call doc will then change to disapprove a PR with such a description added. Running the mapping rules backwards is ambiguous - but it does not matter. The mapping rules are a many-to-many relationship that define equivalence classes of abstract syntax ontolofies and RDF graphs such that the equivalence classes are in a one-to-one mapping and every member of each equivalence class means the same as every member of the related equivalence class. Jeremy
Received on Thursday, 27 March 2003 09:39:09 UTC