OWL (Web Ontology Language)

OWL, the Web Ontology Language, is a W3C standard for describing ontologies on top of RDF. The acronym is deliberately scrambled (after the Winnie-the-Pooh character Owl) rather than written WOL. OWL was first issued as a W3C recommendation in February 2004, and was substantially revised as OWL 2 in October 2009. OWL extends RDF Schema with the constructs needed for formal ontology engineering: named classes and class expressions (intersection, union, complement); property characteristics such as transitive, symmetric, functional, and inverse-functional; cardinality restrictions; disjointness and equivalence between classes and individuals; and value restrictions on properties. These constructs have a formal model-theoretic semantics, which is what enables reasoners to infer entailed triples that were not asserted directly. OWL 2 defines three computational profiles, each trading expressiveness for tractable reasoning. OWL 2 EL is optimized for very large ontologies of mostly subsumption axioms, such as the SNOMED CT medical terminology. OWL 2 QL is designed for query answering over large instance data backed by relational databases. OWL 2 RL is designed to be implementable on top of rule engines. A fourth flavour, OWL 2 Full, is unrestricted but undecidable. OWL ontologies are commonly authored in tools such as Protégé (Stanford) and serialized in Turtle, RDF/XML, or the OWL/XML format. Major deployments include biomedical ontologies (Gene Ontology, SNOMED CT, the OBO Foundry), the schema.org backbone (in a simplified form), and several library and cultural-heritage vocabularies. OWL is a core layer of the Linked Data and RDF stack, although in practice many linked data publishers stop at RDFS and SKOS and never use the full expressive power of OWL.