Re: shapes-ISSUE-86 (dimitris): Associating shapes with ontologies or vocabularies [SHACL Spec]

On 9/11/2015 16:10, Dimitris Kontokostas wrote:
>
>
> On Fri, Sep 11, 2015 at 3:27 AM, Holger Knublauch 
> <holger@topquadrant.com <mailto:holger@topquadrant.com>> wrote:
>
>     One design pattern that will emerge is to have 3 graphs:
>
>     - dataGraph owl:imports ontology ; sh:shapesGraph shapesGraph .
>     - shapesGraph owl:imports ontology
>
>     rdfs:subClassOf triples are needed to be present both in the
>     shapes graph and the data graph in several places, e.g. to
>     validate sh:valueClass constraints. Yet the shapes also need the
>     ontology for the sh:scopeClass triples.
>
>
> I agree that the ontology should exist along with the data but not 
> necessarily together with the shapes. We don't need to perform any 
> transitive closure of rdfs:subClassOf in the scope, or are we?
>

Many shapes will want to use sh:scopeClass, and for that to work best 
(e.g. in editing tools) it is good to have the classes handy. Shapes 
will also reference properties from other namespaces, and it will be a 
good practice to owl:import those. I am not aware of the need for 
transitive class traversal in the shapes graph.

Holger