OWL Web Ontology Language

What OWL is

OWL, the Web Ontology Language, is a W3C Recommendation dated 10 February 2004^[1]. The OWL Reference describes it as a semantic markup language for publishing and sharing ontologies on the World Wide Web, and as a vocabulary extension of RDF^[2].

An ontology, in this sense, is a written-down agreement about the terms of a field: which kinds of things exist in it, what properties they can have, and how they relate to one another. Once that agreement is written in a machine-readable form, different programs can use the same definitions instead of each hard-coding its own.

Why it exists

W3C says OWL is designed for applications that need to process the content of information instead of just presenting information to humans^[1]. A page written for a reader says what should appear on screen; a document written in OWL says what the terms mean, so that software can act on them.

W3C says OWL facilitates greater machine interpretability of Web content than XML, RDF and RDF Schema by providing additional vocabulary along with a formal semantics^[1]. The earlier layers each do part of the job: XML fixes the structure of a document, RDF writes information as subject–predicate–object statements, and RDF Schema adds classes and properties so those statements can be grouped. What those layers lack is a precise account of what the terms imply. A formal semantics supplies exactly that: every construct in the language has a meaning defined precisely enough that a program can draw conclusions from a set of statements rather than merely store them.

Three levels

W3C defines three increasingly expressive sublanguages: OWL Lite, OWL DL and OWL Full^[1]. Each level lets an author say more than the one before it. Publishing the language in layers means a project can write its ontology at the level that covers what it needs, rather than always using the most elaborate form. The middle name, DL, refers to description logic, the family of formalisms whose style of definition the language follows.

A standard in several documents

The OWL Overview says it is one of six parts of the W3C Recommendation, and that the normative formal definition is in OWL Semantics and Abstract Syntax^[1]. So OWL was not published as a single text but as a package: an overview for orientation, a reference describing the vocabulary, and a separate document that carries the formal definition on which the others depend. Splitting the work this way keeps the readable description of the language apart from the mathematical definition that makes it precise.

Where it came from

OWL is derived from the DAML+OIL Web Ontology Language^[2]. The OWL Reference lists its editors as Mike Dean of BBN Technologies and Guus Schreiber of the Free University Amsterdam^[2], and it lists authors from the University of Manchester, the Free University Amsterdam, the University of Maryland, Stanford University, Bell Labs Research and Franklin W. Olin College of Engineering^[2]. The spread of institutions — European and American, industrial and academic — reflects the kind of collaboration a Web-wide vocabulary standard requires: the definitions only work if everyone agrees to use the same ones.

What it changed

OWL's contribution is a division of labour on the Web. RDF already gave data a common sentence form; OWL gave it a common way to state what the words in those sentences mean, and to do so with a formal semantics that software can reason over^[1][2]. That is what makes it possible to publish an ontology at a stable address and have other people's applications use it, instead of every system keeping a private dictionary. OWL's place in the standards stack is therefore just above RDF, whose vocabulary it extends^[2], and its design point — richer vocabulary plus formal meaning — is the thing later work in the same area has had to reckon with.