- From: Michael Kifer <kifer@cs.sunysb.edu>
- Date: Wed, 24 Aug 2005 12:50:08 -0400
- To: Jim Hendler <hendler@cs.umd.edu>
- Cc: Dieter Fensel <dieter.fensel@deri.org>, public-rule-workshop-discuss@w3.org
Jim Hendler wrote: > > Dieter wrote: > > > >On the other hand, it is not that difficult to identify something > >around Horn logic > >with a minimal model as common ground of state of the art rule languages. > > > Mike wrote: > > >The only rule language that FOL is a superset of (modulo a mapping) is SWRL > >- an untested newcomer. And of Horn Datalog, which is near-useless. All the > >real rule-based languages don't map into FOL AFAIK. > > Am I the only one confused? Why is "Horn logic with a minimal model" > so preferable to "Horn Datalog"? Jim, I didn't write about the minimal models. I was trying to stick to the precise definition of "Horn", which doesn't include negative queries. So, in this case, minimal models are not needed. Dieter probably had negative queries in mind. But, as I wrote before, pure Horn is near-useless, so in the end we need negative body literals, minimal models, etc. --michael
Received on Wednesday, 24 August 2005 16:50:17 UTC