Re: [topbraid-users] right owl2shacl-mapping?

2021-08-31 Thread Holger Knublauch
This (interesting) design may work technically, but it can be very slow because it will traverse all rdf:type triples everywhere and use a rather complex algorithm with nested (qualified) shapes. A better solution would indeed start with exactly the instances of one of those classes. The

RE: [topbraid-users] right owl2shacl-mapping?

2021-08-31 Thread 'Bohms, H.M. (Michel)' via TopBraid Suite Users
Hi David Your solution (as generated) needs quite some code in case of more than two items and also in a distributed way. We hoped to cover it with more generic template code… Wrt shacl AF: if we can do it in shacl core (being a recommendation) that would be preferred. Furthermore I would not

Re: [topbraid-users] right owl2shacl-mapping?

2021-08-31 Thread David Price
I don’t like creating a separate instance to control disjointedness as it’s not immediately visible in the relevant class form. The NodeShape being proposed has the same issue. I usually use a pattern like this: unnamed:Thing_1 rdf:type owl:Class ; rdfs:subClassOf owl:Thing ;

[topbraid-users] right owl2shacl-mapping?

2021-08-31 Thread 'Bohms, H.M. (Michel)' via TopBraid Suite Users
(we were looking for an approach not needing rdfs-entailment) sml:AllDisjointClasses_1 a owl:AllDisjointClasses ; owl:members ( sml:PhysicalObject sml:InformationObject sml:Activity sml:Event sml:State ) ; . In shacl (?): sml:DisjointClassesShape_1 a