Entry level OWL inferencing becoming more and more feasible…

I wonder whether it would be different if OWL 2 RL was used instead of OWL Horst (i.e., to see if OWL 2 RL introduces features that make the possible optimizations difficult or impossible…). This would be important to know, if the community would define some sort of a sub-set of OWL 2 RL for practical purposes. Or is OWL Horst really that subset (there are features in OWL 2 RL that are really cool, like property chains or keys…)

See you guys around in Bonn to discuss this further…