- From: Holger Knublauch <holger@topquadrant.com>
- Date: Thu, 18 Dec 2014 09:41:03 +1000
- To: public-hydra@w3.org
On 12/18/2014 9:22, Markus Lanthaler wrote: > What exactly do you mean by global range/domain? The rdfs:domain (and schema:domainIncludes) statements are global in a sense that they do not depend on the context of a class. >> 2) Define an abstract superclass and use inheritance >> >> schema:email rdfs:domain schema:ThingWithEmail >> schema:Person rdfs:subClassOf schema:ThingWithEmail >> schema:Organization rdfs:subClassOf schema:ThingWithEmail > Yeah.. that's the "traditional" way of doing it.. it normally introduces > quite a lot of "artificial" classes which are only there due to those > modeling constraints and I'm not entirely sure what you gain by that You gain that you don't need a notion of global constraints, and properties no longer have to be stand-alone entities. This means that you have better interoperability between RDF and other languages such as OO systems (and even XML) where all properties are locally scoped to classes. This in turn reduces the learning curve, creates more direct mappings and round-tripping (e.g. think about all those Java class generators from ontologies). > And what prevents you from doing that with global properties? Nothing, but why invent a new concept (global properties) that is not supported anywhere else than in RDF? Global constraints are too limiting in their expressivity, e.g. they don't easily allow you to reuse the same property in different classes (see current threads on the schema.org mailing list). Furthermore, by always creating superclasses you just need a single mechanism to understand the constraints: just walk up the class hierarchy. In the current OWL world, you would need to check both for local owl:Restrictions (walking up the class tree) *and* look for global rdfs:ranges. This makes all algorithms more complicating. And in many (if not most) cases, a property has a single natural class to attach to, so my example of introducing an abstract superclass for emails was already quite complex. Regards, Holger
Received on Wednesday, 17 December 2014 23:44:08 UTC