search

w3c / dx-prof

The Profiles Vocabulary
https://w3c.github.io/dx-prof/prof/
Other
5 stars 2 forks source link

Alignment / differentiation with OWL #10

Open nicholascar opened 5 years ago

nicholascar commented 5 years ago

Raised by Reviewer 1 of the Profiles Ontology ESWC paper and elsewhere: PROF must differentiate itself from OWL, specifically prof:isProfileOf v. owl:imports.

nicholascar commented 5 years ago

Closing since we see no need to include a specific handling of PROF / OWL in doc

nicholascar commented 5 years ago

While recognising that we've stated "we see no need to include a specific handling of PROF / OWL in doc", we have added an OWL section to "Related Work" for the 3PWD candidate so others thinking about PROF/OWL relations will have our thoughts there, see https://raw.githack.com/w3c/dxwg/prof-3PWD-candidate/prof/index.html#related-owl

aisaac commented 5 years ago

I disagree with closing this issue.

There is a section in the current draft, but it does not mention prof:isProfileOf (contrasting it with owl:imports) as requested in the issue description. (for the record I think we could say that owl:import is a sub-property of prof:isProfileOf, but maybe more discussion is needed).

I also disagree with the content of the section: especially the paragraph on owl:imports that leads to the conclusion "It is not possbile [note typo here!], with owl:imports, or any other OWL ontology property, to know that data conforming to an ontology also conforms to another ontology.". Having an statement X owl:imports Y means that every class, property and axiom accessible in the ontology (document) Y is included in X. If X contains all axioms of Y, this means that any data conforming to X will automatically conform to Y (since the import does not 'de-activate' specific axioms/constraints of the imported ontologies).

Finally the section does not mention OWL Profiles and I think anyone knowing OWL and seeing a section on OWL in a document about profiles will be quite badly surprised if they don't see OWL Profiles being discussed.

nicholascar commented 5 years ago

Removing "due-for-closing" since we have unresolved issues.

nicholascar commented 4 years ago

I agree with @aisaac's characterisation of the owl:imports / prof:isProfileOf relationship (that we should have owl:imports rdfs:subPropertyOf prof:isProfileOf.

I've changed the paragraph @aisaac took issue with in the Related Work section (not PRed through yet though) to reflect this thinking to:

Within OWL, the definition of owl:imports is not associated with any axioms that permit reasoning with it and it is mostly used by ontology editing software to manipulate ontology data by importing it from remote (online) locations, as a technical convenience mechanism. Since an ontology that imports another includes all the axioms of the imported ontology, data which conforms to an ontology x will necessarily conform to any ontology that x imports too.

When using OWL (and RDF [RDF11-CONCEPTS] & RDF Schema [RDF-SCHEMA]) it is possible to know, on a per-class or property basis, that data elements conform to particular ontologies through inspection of those elements. Data that uses OWL declares classes and properties as being of types – particular classes or properties – by using rdf:type – :myNewClassInstance rdf:type ex:SomeExistingClass – which claims the data conforms to the indicated class or property. Newly declared classes and properties in ontologies may also claim conformance to existing classes or properties by using rdfs:subClassOf or rdfs:subPropertyOf respectively.

I'd like more discussion about owl:imports rdfs:subPropertyOf prof:isProfileOf, before adding it to the ontology, @rob-metalinkage ?