Back when I first saw the Description Logic-based standard for OWL, I wondered if the German logicians who had developed things like Frame Logic and KARL were asleep at the wheel when the standard was completed. They have moved forward in their work, with the success of companies like Ontoprise, but they had to deal with the fact that the logic they use is not conformant to any of the W3C Semantic Web standards. For a number of reasons I will outline below, I wondered if this would be the year that the continent would strike back, and try to repair this damage.
So I told some of my students that I expected there to be a demonstration at ISWC 05 of the supporters of a datalog version of OWL (so-called "OWL-DLP") as a response to the current description logic version of OWL ("OWL-DL"). This was based in part on provocative paper titles like DLP isn't so bad after all, as well as papers like this one that work out the formal background for a datalog based version of OWL, and KAON-2, a datalog system that does a very good approximation to OWL. The possibility has been discussed by a variety of unlikely authors (including Horrocks, Parsia, Patel-Schneider and Hendler), who also warn of the ramifications of such a thing. I think that all of this is partly in response to a growing feeling that maybe the value that description logic provides isn't necessarily the value that all semantic web users need.
Well, there wasn't a big demonstration at ISWC (or at the OWL workshop that followed), but something did happen at ISWC that has set some wheels in motion. In his keynote talk (esp. slide 12), Tim Berners-Lee suggested that there was a place for OWL-DLP in the famous "layer cake" tech rollout plan for the Semantic Web. Ontoprise, whose OntoBroker is based on a datalog reasoner, sees this as a breakthrough in viewing OWL-DLP as a valid and standard-worthy part of the Semantic Web. Perhaps it is the drama queen in me who wanted to see a knock-down drag-out fight, but it is much more realistic to expect what has actually happened; one datalog vendor has simply found a reason to declare victory, and find their place in the new layer cake. And perhaps this is as it should be; there are those who find distinct value in the Description Logic foundation of the Semantic Web, and there are those who find value in a more procedural description. As Sir Tim said, there's room for a lot of languages in the Semantic Web.
