OWL2-RL rules in Thea
Chris Mungall
http://www.edbt.org/Proceedings/2011-Uppsala/workshops_toc.html
The author (cc'd) says that there's no RL engine in Thea - it's quite possible that Vangelis' engine did not exist when Jesus wrote the paper.. or maybe we just need to make it more prominent!
The paper has a nice detailed description of the algorithm used (which we lack in Thea!). But from a cursory reading they sound similar. The paper describes an implementation directly over triples, whereas Thea works over OWL2 facts. I wonder if this makes a difference to performance. It would be interesting to compare the two.
Perhaps we could combine Jesus' lib (with permission) into Thea and give users a choice - although moving forward it may make sense to have a single RL implementation for SWI if they're using essentially the same forward chaining algorithm.
Cheers
Chris
Vangelis Vassiliadis
Indeed we did not make any publication for the RL reasoning module in Thea. This is because mainly
due to my (very) limited time and lack of resources in general.
If Jesus would like to work in this direction, i.e include his RL engine into Thea or investigate merging the two libraries I'd be more than happy
to help him.
Kind regards,
Vangelis