Despite various gross ontology engineering methods, philosophical principles and work on design patterns, there is no work that informs us about the ontology authoring process at the level of addition of axioms. Logic based ontologies, such as those written in OWL, can be complex systems of axioms where a single change can have ramifications across the ontology.
Authors can call upon a reasoner to ‘check’ their ontology and re-organise it according to the implications of the added axioms. This is, however, a post hoc activity; an ontology author lacks the means of undertaking ‘test driven development’ (apart from issuing queries against the reasoned ontology). Our hypothesis is that by allowing an author to pose ‘what if…?’ questions to an ontology prior to the addition of axioms, then the authoring process will run in a more informed manner. In order to test this view, we need to understand ontology authoring at the axiom level to find out how users decide upon the addition of axioms; that is, the dialogue they have with the system.
Software
Protégé4US – Protégé for User Studies (version 0.1, available soon)