Ontologies that provide shared domain theories will be a key asset for knowledge management in general and particularly for the Semantic Web. They could encode both "trivial" common-sense and expert knowledge which is a prerequisite for "intelligence" - the ability of a person or machine to adapt and behave adequately in diverse and complex circumstances. An ontology is the basic knowledge about the world (or just a domain) that is reusable because it is independent from a particular system, task or situation. Ontologies represent the "pre-compiled" knowledge, the semantic reservoir that enables people to communicate and reason efficiently on top of few (task- or situation-) specific facts. For instance, a general ontology could contain information such as the fact that "ducks are birds" - humans do not infer this each time they think about duck, they know it! This fact (together with a lot more information about the properties of birds and how the ducks are different and typical and so on) is part of people's ontology about the world.

Note: there are number of alternative definitions of "ontology". We are only going to mention the basic one that comes from the philosophy: Ontology is the science about the nature of being and existence.

From a theoretical viewpoint, each document on the web can be considered as consisting of two bodies of information:

• information about a state of affairs involving some instances (or individuals) and relations among them and
• an ontology reflecting the general knowledge related to this state of affairs, about the instances participating in it and how they are situated within a more general ontology.