Re: OWL URI and namespace, an issue for LOV.

On Tue, Jun 17, 2014 at 1:00 PM, Bernard Vatant <bernard.vatant@mondeca.com>
wrote:

> Hi Simon
>
>> It is forbidden to treat blah# and blah as being sameAs without an
>> explicit axiom; i *think* it is an error if ontology is named blah, but is
>> imported as blah#.
>>
> "is imported", you mean in LOV data base for example?
> See previous answer to Antoine. We do as we can with what we get :)
>

I was referring to OWL Import statements (see: <
http://www.w3.org/TR/owl2-syntax/#Imports> ).

Ontology Documents are defined in <
http://www.w3.org/TR/owl2-syntax/#Ontology_Documents>.

Namespaces  and documents don't exist in OWL once it's been processed.
 It's all one big cauldron of axioms.

It's possible, e.g., to define  dc properties in an unrelated ontology,
without implying that the dcterms ontology is imported.

Redirection can happen, but the results must be indistinguishable from
direct retrieval of the original.

If an imported ontology has no Ontology IRI defined, then the imported
ontology is anonymous;  if the ID is the empty fragment, the it resolves to
the base, which should be the requested IRI (otherwise it would be
distinguishable).

The default OWL API ontology manager does not check to see if there is a
mismatch between the document IRI and the (which I think is for error
tolerance).  An ontology with a matching ontology or version IRI will still
be preferred.

Don't get me started on VersionsIRIs...