A bit ontology

The term 'Ontology' is being used here in the meaning of Web Ontology Language (OWL). With OWL computer applications are able to 'understand' the meaning of a concept. In addition with OWL 'reasoning' is possible, which means that correlations between concepts can be found that are not explicitly defined, reasoning is executed by special applications. An example of such a correlation is that the brother of my mother is my uncle. 'Understanding' a concept implies that the meaning of a concept is made explicit. In its simplest form a concept is made explicit by attaching a URI (Unique Resource Identifier) to the concept. A concept with a URI can be used in (computer) communication to ensure that a subject is one and the same concept. This does not give a meaning to the concept yet, but it does provide a context in which the concept is being used. The reference to OWL above, for example, provides a context for 'Ontology', which otherwise could be understood differently. Meaning can be added to a concept through allocated properties, the set of properties forms the definition of the concept.

An OWL file basically consists of a series of statements, approachable through a URI, the URI mostly being a web-address. A statement consists of three parts: subject - predicate - object, known as a 'triple'. Instead of 'predicate' the term 'property' is used, here we us predicate for the correlation between subject and object to prevent confusion, whereas property is used for a complete statement.

OWL distighuishes between classes and individuals (or instances), where classes are generalizations of individuals, the distinction between the two is not not always clear, however. Individuals always refer to the class(es) of which they are derived. Classes can be organized in a class hierarchy of super- and subtypes. Subtypes and individuals inherit the properties of their supertypes.

In OWL there are two types of predicates: 'objectproperties' and 'datatypeproperties'. Objectproperties are basically concerned with correlations between classes or individuals derived from these, datatypeproperties describe properties that may have a value. An example of statement with an objectproperty is: 'Building floor' - 'is a part of' - 'Building', an example of statement with a datatypeproperty is: 'Beam' - 'has length (mm)' - '3000' . In addition to these two types of predicates informal statements (annotations) exist, such as labels, comments and references.

As with databases an ontology can be 'queried' and edited by means of a query lanmguage, such as Sparql.

2015-01-15