- From: Richard Cyganiak <richard@cyganiak.de>
- Date: Tue, 31 May 2011 19:52:05 +0100
- To: Juan Sequeda <juanfederico@gmail.com>
- Cc: ashok.malhotra@oracle.com, RDB2RDF WG <public-rdb2rdf-wg@w3.org>
Juan, On 31 May 2011, at 19:35, Juan Sequeda wrote: > and rdfs:range triples. Why are they necessary for reversibility? > Would it be necessary to have rdfs:Class/owl:Class and rdf:Property/owl:ObjectProperty - owl:DatatypeProperty ? I assume no, because this can be inferred. But adding them wouldn't be a harm (right?). If so, we are also specifying a direct mapping of the relational schema to RDFS/OWL. Careful. I think I've been saying this to you quite a few times over the course of this WG: If you claim to be mapping the schema, then you have to map it properly. There are SQL datatypes, nullability, multi-column primary keys, foreign key constraints, unique constraints, the works. I'm not sure that all of this *can* even be expressed in RDFS/OWL. And I doubt even more that it is useful. Personally I would prefer if the direct mapping graph was not littered with OWL statements that don't enable any useful inferences. Best, Richard
Received on Tuesday, 31 May 2011 18:52:34 UTC