Best Practice: Be careful with the use of inferencing in RDF Vocabularies

State: Draft

Contact: Nick Crossley

Scope

This Best Practice recommends appropriate usage of inferencing properties such as rdfs:domain, rdfs:range, rdfs:subClassOf, and rdfs:subPropertyOf in RDF vocabularies.

Recommendation

RDFS contains terms such as rdfs:domain, rdfs:range, rdfs:subClassOf, and rdfs:subPropertyOf, which go beyond mere vocabulary definition and enter into the world of ontologies. The primary difference between a vocabulary and an ontology is that an ontology includes inference rules which let you infer new information from given information - that is, add new triples to those stated, rather than constrain data to some schema.

For more information on this, read OSLC Resource Shape - A language for defining constraints on Linked Data.

In the light of this knowledge, be careful with the use of inferencing terms in vocabularies - they often do not mean what you might have thought they meant. In particular, do not use rdfs:domain or rdfs:range with multiple values when defining a vocabulary term for a property, since that infers the target of that property is of all mentioned types, not any one of them.

Examples

Here is an example snippet of an RDF vocabulary document:

@prefix oslc: <http://open-services.net/ns/core#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .

oslc:ServiceProvider a rdfs:Class ;
   rdfs:isDefinedBy oslc: ;
   rdfs:label "ServiceProvider" ;
   rdfs:comment "The Service Provider resource" ;
   rdfs:seeAlso <http://open-services.net/bin/view/Main/OslcCoreSpecification#Resource_Service_Provider> .

See Also

• RDF Schema
• OSLC Resource Shape - A language for defining constraints on Linked Data