- From: Ian Horrocks <horrocks@cs.man.ac.uk>
- Date: Thu, 25 Apr 2002 08:59:23 +0100 (BST)
- To: Pat Hayes <phayes@ai.uwf.edu>
- Cc: www-webont-wg@w3.org
On April 8, Pat Hayes writes: > >On March 21, Dan Connolly writes: > >> On Thu, 2002-03-21 at 14:28, Ian Horrocks wrote: > >> > On March 21, Libby Miller writes: > >> > > > > >> > > > As noted in the design discussions for DAML+OIL, I don't > >> > > > see sufficient justification for making V disjoint > >> > > > from R. > >> > > > > >> > > > It seems silly not to be able to talk about the intersection > >> > > > of two sets of strings, or UniqueProperty's whose > >> > > > range is dates, or whatever. > >> > > >> > This means that any OWL reasoner has to take on responsibility for > >> > reasoning about types > >> > >> I gather when you say "OWL reasoner" you mean a complete > >> reasoner. > >> > >> I'm not very interested in such a thing. > >> > >> Regular old horn-clause/datalog reasoners > >> (with some built-in predicates like > >> string:lessThan and such) seem > >> to get me what I need pretty well. > > > >Dan, > > > >It seems that, on the basis of a few toy examples where using ad-hoc > >reasoning seems give the results you want/expect, you conclude that > >this will be appropriate/adequate for all applications. I don't find > >this argument very convincing. > > > >Even w.r.t. ontology level reasoning I expect things to rapidly get > >large and complex enough that humans wont be able to check all > >inferences - we will just have to trust that the reasoner got it > >right. Soundness is therefore essential, > > True, but > > >and completeness highly > >desirable. > > That is much less clear. Can you spell out the actual argument for > completeness here? I have argued very strongly in the past for > completeness, but that was in a different context: Krep in AI, where > completeness provides a methodological security against > overconfidence in thinking that ones representation captures more > content than it really does capture. I don't see that issue as being > at all central in applied ontology work, particularly on the web. > > >For example, when multiple processes are interacting, some > >action may be taken by one process on the basis of a non-inference by > >another process > > That is a nonmonotonic process, which can lead to unsoundness even > with a complete reasoner. So this seems irrelevant. > > >, so incompleteness can easily lead to "unsoundness". > > > >As far as the disjointness of object/data domains and properties is > >concerned, there are also good pragmatic reasons for this, including > >the ability to use hybrid designs for OWL reasoners, i.e., the ability > >for an owl "object class" reasoner implementation to "bolt on" a type > >checker for arbitrary type systems. > > But such reasoners could still be USED: the use of the more liberal > syntax does not make such reasoners illegal, only possibly less > universally effective. The real issue is whether the web content > language needs to be protected against the risk that it might be able > to express something that many reasoners could not fully utilize. I > would suggest that as a very basic web-methodological point, we > should NOT take such worries as being central concerns in the design > of OWL. The whole web context means that we cannot forsee the kinds > of inference engine that are going to be used, and we should not let > the content language be limited by the needs of the engines that we > happen to know and love at present. It may well be the case that > particular areas of usage will discover, and utilize, combinations of > features which allow for pragmatically useful reasoners that work in > syntactic subclasses that cut across our current implementation > experience. To reiterate, my suggestion is only that we design the language in such a way that it is possible to design/implement effective decision procedures. Of course the definition of "effective" could/will be contentions. Allowing mixed domains will certainly make life more difficult for most system implementors as their reasoners may need a fairly detailed understanding of the characteristics of the domain in question. For example, it would be quite easy to assert that every element of the domain of discourse is a value in some datatype. In this case a reasoner may have to consider interactions between the cardinality of the datatype and models of the abstract domain. Ian > > Pat > > -- > --------------------------------------------------------------------- > IHMC (850)434 8903 home > 40 South Alcaniz St. (850)202 4416 office > Pensacola, FL 32501 (850)202 4440 fax > phayes@ai.uwf.edu > http://www.coginst.uwf.edu/~phayes >
Received on Thursday, 25 April 2002 10:36:57 UTC