Scaling OWL (two new ways)
by Bijan Parsia
For those with OWL-performance-fear, there is some good news. There usually is, but these tidbits are striking:
- On the TBox side of things: Boris Motik and Rob Shearer (with Ian Horrocks) have developed a new reasoning calculus that is very effective with the notorious Galen ontology, and, indeed, with all the OBO ontologies. They tackle both non-determinism and tableau size with stunning results. It should also have positive implications for DL Safe rules. They have a prototype reasoner using the technique, HermiT, available for download.
- On the ABox side of things: IBM Research (Watson Research Center, NY) have recently posted information about their summarization technique for scaling ABoxes. They have a reasoner, SHER, which will be available in one form or another at some point (note! it has Pellet Inside!). Their case study is quite inspiring.
(The hypertableaux calculus has positive implications for ABox scaling as well.)
Since I’m an hour train ride from Liverpool, I’ll just conclude that it’s getting better all the time. Read the papers. Enjoy.
