[JENA-30] Inconsistency in rdfs8

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: Reasoners
• Assignee: Dave Reynolds
• Reporter: Dave Reynolds
• Created: 18/Jan/11 13:56
• Updated: 18/Jan/11 13:56

Description

There is an inconsistency between OWL_MICRO and OWL_MINI/OWL_RULES in handling of subClassOf hierarchies without class declarations.

Model model = ModelFactory.createDefaultModel();

String NS = "http://www.openjena.org/tests#";

Resource Person = model.createResource(NS + "Person");
Resource Mammal = model.createResource(NS + "Mammal");
Resource Animal = model.createResource(NS + "Animal");

model.add( Person, RDFS.subClassOf, Mammal);
model.add( Mammal, RDFS.subClassOf, Animal);

// Without these the class hiearchy closure is incorrect in Mini/Full
model.add( Person, RDF.type, OWL.Class);
model.add( Mammal, RDF.type, OWL.Class);
model.add( Animal, RDF.type, OWL.Class);

/create an OWL inference model based on previously created model./
InfModel inf = ModelFactory.createInfModel(ReasonerRegistry.getOWLMiniReasoner(), model);
printStatements(inf, Person, RDFS.subClassOf, null);

The problem is that that indirect deduction of rdfs8 via prototypical reasoning relies on owl:Class whereas type inference will only deduce rdfs:Class