Ontologies define the ways we imagine reality. We form ideas, concepts and imaginations that we believe reflect the reality or some reality. While in philosophy there is just one ontology per system computer science has no problem in creating several ontologies. These usually cover only a part of the world and human activity, while philosophical ontologies are usually holistic.
Ontologies only make sense if they are shared. Various parties must refer to the same set of concepts to be able to communicate. Presently there are only few ontologies that are public and shared. Some of them in the biological-medical area (Gene Ontology to search PubMed), one that is created by the big search engines (Google, Yahoo, Microsoft, and Yandex) and Good Relations for e-commerce. But even within some project non-public ontologies make sense. They can enhance search and data mining functions. So some travel-sites use internal ontologies to make it possible to search by natural language phrases (e.g. Fact-Finder of Omikron Data in www.neckermann-urlaubswelt.de or www.gulet.at ).
In the realm of computers there are special languages that are used to define ontologies. Today usually RDFS (Resource Description Framework Schema) or OWL (Web Ontology Language) are used. Semantic techniques are not confined to the use of one ontology. Several ontologies, public and non-public, can be used if the relation between them is ascertained. Ontologies can mediate between several data structures and even unstructured data.
Established Ontologies in the Web:
- Good Relations – Ontology for e-commerce, about to be integrated into schema.org
- Accommodation Ontology – extension of Good Relations for Accommodations (Hotels, B2B etc.)
- Gene Ontology – major bioinformatics initiative to unify the representation of gene and gene product attributes across all species
- schema.org – Reference-System for Search Engines, probably not an ontology in the full sense
Some philosophical thoughts about agile ontologies (in German) at: