- From: Bijan Parsia <bparsia@isr.umd.edu>
- Date: Fri, 20 May 2005 07:14:04 +0900
- To: Daniel Elenius <daele@ida.liu.se>
- Cc: public-sws-ig@w3.org
On May 20, 2005, at 2:29 AM, Daniel Elenius wrote: >> Rules in SWRL are just material conditionals. They most certainly >> "evaluate" to true or false. They may be inferred. > Hmm, ok, I take that back then. > >>> In fact, it may be easier to just do AtomLists (if there's a single >>> Atom, then we just have a list with one element). >> >> >> Prolly true. But if Preconditions are just formulae (rather than >> restricted to conjunctions), there's no need to avoid imps. >> > I guess the only reason would be that the any OWL-S tools (not just > editing tools, but especially execution engines/planners) would need to > understand Imps, whereas if just AtomLists are allowed, things are > somewhat easier. Imps make things undecidable, right? Hmm. Well, my guess is that OWL-S is undecidable for many reasons :) Making it OWL-Full (as removing quotation will do) might not *alone* do it, but it's prolly bad. > (Whereas AtomLists > (with instantiated parameters) correspond to ABox queries.) Hmm. Don't know that that's decidable if you allow undistinguished variables. Interesting question > That may not > be a reason to prohibit them though. Yeah. A reason would be that such universal quantifications are just weird as a precondition. The counter would be, "So they're weird! Who cares? Let the chips fall!" We certainly allow arbitrary KIF formulas. Cheers, Bijan.
Received on Thursday, 19 May 2005 22:14:11 UTC