0

I'm using HermiT 1.3.8.3 to reason over an OWL ontology (230 classes, 21 object properties, 30 instances and 30 SWRL rules) both inside Protege (4.1 and 5.0) and in a Java application I'm developing using OWL API 3.5.0.

The problem is that while in Protege HermiT takes 225 seconds to classify an realize the ontology, in my application it spents twiceor three times as much time to do the task with the very same ontolgy. Besides that, any query that DL-query tab of Protege answers in a few seconds takes minutes to be answered inside the application (using the technique of creating a OWLClassExpression equivalent to the query and making HermiT reason over it).

Does anybody know why it is happening? Is it a matter of reasoner configuration or something like that?

Thank you!

Joshua Taylor
  • 84,998
  • 9
  • 154
  • 353
TitoFHR
  • 87
  • 2
  • 8
  • Are they the same version of HermiT? – Joshua Taylor Aug 24 '14 at 18:16
  • I think so. The one I used inside the application I downloaded from http://www.hermit-reasoner.com/download/1.3.8/. In Protegé I have both 1.3.8 and 1.3.8.3. Anyway, I also tried to put inside the application the very same .jar that comes as Protege plugin - what makes it even more weird... – TitoFHR Aug 24 '14 at 23:15
  • duplicated at http://answers.semanticweb.com//questions/29926/why-is-hermit-slower-in-stand-alone-java-application-than-in-protege – Ignazio Aug 25 '14 at 10:16
  • Can you show us the code you're using to call HermiT and the ontology? The only configuration that should make a difference is the amount of memory available for the heap - are you passing any value as -Xmx at runtime? Protege (both 4.1 and 5) should not be doing anything fancy with HermiT, so I'd expect the speed to be the same in both cases. – Ignazio Aug 25 '14 at 10:17

0 Answers0