The most popular approach of creating ontologies is engineering-oriented, i.e., a small number of individuals carefully constructs the representation of the domain of discourse, and releases the results at some point in time to a wider community of users. The sequential paradigm of this approach, and the fact that a small group constructs the ontology for a bigger group has several weaknesses.
First, the ontology evolution is not under the full control of the ontology user community. For example, missing entries cannot be added by any user who reveals the need for a new concept, but has to be added by the small group of creators. This slows down ontology evolution and there is very likely a loss of user feedback. In comparison, in natural language, the evolution of the vocabulary is under the control of the user community. Speakers can introduce new and refine or confirm existing entries in the vocabulary.
A second weakness of is that users creating annotations cannot easily grasp the intension of a concept. There is a lack of communication between the ontology creator and the broader user community.
In the MyOntology project, we propose to use the infrastructure and culture of Wikis as an ontology workbench that fosters true collaborative, community-driven ontology creation and maintenance. We especially propose the use of multimedia elements to improve the expressiveness and disambiguity of informal concept definitions in an ontology. Also, we regard it as beneficial if the definition of a concept is not separated from the discussion that lead to shaping the intension of this concept.
In order to achieve this goal, we will (1) develop an ontology environment based on Wiki technology and a scalable ontology repository, (2) analyse the processes and roles in achieving consensus about a domain of discourse, (3) map the various types of functionality available to a user (e.g. adding a named class vs. adding a relation), to proper mechanisms that help create useful ontologies (e.g. based on voting, explicit endorsement, usage analysis, or other innovative techniques of social software), (4) develop small, yet effective functionality that supports users on one the hand, and makes abuse or malicious behaviour unattractive on the other, (5) develop cognitively adequate user interface (UI) techniques, and (7) validate our approach by setting up a test platform and and populating an ontology.