T. R. Gruber. Toward principles for the design of ontologies used for knowledge sharing. In N. Guarino and R. Poli, editors, Formal Ontology in Conceptual Analysis and Knnowledge Representation, Padova, Italy, March 1993. International Workshop on Formal Ontology, Kluwer Academic Publishers.
This paper describes some design principles useful when designing an ontology. The starting argument is that ontologies can play a specification role in the establishments of agreements among communication peers. Ontologies are specification mechanisms. Knowledge is based on some conceptualization which are simplified, abstract, representation of the world. Therefore, an ontology is an explicit specification of a conceptualisation. Two interesting ideas emerge from the paper:
1) an agent commits to an ontology if his observable actions are consistent with the definition of the ontology: the implicit idea is that the agent has to follow 100 per cent the specification of the ontology, whereas there may be the case in which the agent is following only partially the ontology specification. Can we talk about progressive commitment?
2) Knowledge is attributed to agents observing their actions: here a more epistemological consideration can be asked: is knowledge in the community, in the individuals rather than into the system?
The five criteria for the design of an ontology are:
1. Clarity: definitions should be objective;
2. Coherence: defining axioms should be logically consistent;
3. Extendibility: it should anticipate tasks monotonically without the revisions of the existing definitions;
4. Minimal encoding bias: representation choices should be made as less as possible for the convenience of notation;
5. Minimal ontological commitment: the ontological commitment should be sufficient for supporting the intended knowledge sharing.
The author says that ontology design will require making tradeoffs among the criteria. My point is that: should this be part of the collaborative negotiation process?