Evaluating a Temporal Causal Ontology

We describe an experimental method to evaluate options within an ontology. (I) We commit to some precise description of the task of a theory written in an ontology. (II) A mutator automatically generates variants to known problems according to the diierent ontological options. (III) We try to perform the task over the mutations. (IV) We search for cases where the task fails. In this framework, ontological options can be assessed as follows. The better ontological options fail less on the task. This framework is instantiated in the context of a causal ontology. Four variants are proposed for handling time in the given ontology. Two of those four are found to be clearly superior and two were found to be clearly inferior.