I am not sure whether I agree with the second paragraph. It depends how wide you understand "use case" and on our scope of "ontology evaluation". If "ontology evaluation" is supposed to cover the evaluation of every aspect of an ontology that is important to decide which of a given set of candidate ontologies one should choose for an IT system, then there are considerations that are not directly linked to the functioning of the ontology within the IT system. For example, you might choose an ontology that is still maintained by a large community over an ontology that has not been updated since 1995. Or you might choose an ontology that has an "open" license over an ontology which comes with an expensive commercial license.
Of course, if you understand "use case" to cover evaluation criteria like "The ontology needs to be maintained by somebody else" and "The ontology needs to be free, because our project has no money for licenses", then I agree with you. :-)
Fabian
On Dec 12, 2012, at 5:30 PM, Amanda Vizedom wrote: Indeed, we need to keep in mind this point, well-made during last year's summit: even applied, formal ontologies are not always used as part of an IT system. They maybe used as part of a primarily human system, e.g., as a method of building a model of something complex in order to understand it better. And even those that are part of an IT system don't get all -- perhaps not even most -- of their requirements from the IT; many requirements derive from the intended use and its human, business process, and other characteristics.
That said, it is still fair to say that essential requirements -- those from which evaluation dimension relevance is derived -- are themselves derived from the use cases. We just need to avoid artificial closure of the types of use cases. If it is a real ontology (whether or however embedded in IT) use case, it matters to understanding what, and when, ontology evaluation dimensions are important.
Amanda
<ATT00001..c> |