In the previous installment we concentrated on how to encode literature information into a topic map. During this process we had to commit ourselves to particular design decisions. These we will have to adopt consistently in the authoring or in the maintenance phase of a map (or a set of maps).

While an informal understanding may well work in small teams, distributed authoring calls for a more disciplined approach to capture the rules what exactly maps should contain and how they are going to be built. This where ontologies, in our case Topic Map-based ontologies come into play. They behave as the schema which maps should adhere to.

One of the first purposes of an ontology is to define the concepts and to organize them into a type system (taxonomy). In a further step we have to define rules on how topics such as those for documents have to be formed and what information they must or may contain; the same applies to the associations in a conforming map. Finally, we have to exert more global control by expressing some application specific constraints.

This classification of constraints into taxonomy, structure and application specific is somewhat arbitrary but helps to organize ontologies.