Seven golden rules for a web rule language

Web Ontology Language is now the W3C’s candidate recommendation, 1 which makes me think that the promises of the Semantic Web will come closer to being realities.2 Right? A close reading of the famous Scientific American article and comparison with OWL reveals, however, that OWL cannot account for rules such as "If a city code is associated with a state code, and an address uses that city code, then that address has the associated state code." OWL doesn’t even respond to trivial conversion between measures in different systems (say meters versus feet). At this point, the famous Semantic Web layer cake Tim Berners-Lee developed comes in handy (www.w3.org/2001/09/06-ecdl/slide17-0.html). Beyond a solid foundation of RDF, RDFS, and OWL, we are missing rule standardization. This, in turn, triggers the question that seems to be paralyzing further standardization—namely, "Which type of rules do we actually need?" Some favorite types include event-condition-action rules (such as triggers in databases), first-order Horn logic axioms, semantic translation rules, and so on. In this issue, you’ll find a range of answers, from "Do we really need rules?" (Ian Horrocks) to "We need all different type of rules at once!" (Gerd Wagner). Because no canonical answer seems to exist, I guess the promises of the Semantic Web will have to wait. But, let’s think about what the next step should be