- From: Bijan Parsia <bparsia@isr.umd.edu>
- Date: Fri, 25 Jun 2004 13:34:59 -0400
- To: Donal Murtagh <domurtag@cs.tcd.ie>
- Cc: public-sws-ig@w3.org
On Jun 25, 2004, at 11:43 AM, Donal Murtagh wrote: >>> If I recall correctly, you suggested that Pellet's conjunctive ABox >>> query answering module could be used to evaluate such preconditions. >> >> Yes. >> >>> Is >>> this true, and if so, what syntax should they be expressed in if >>> Pellet >>> were to be used for this purpose? >> >> We will support, via the OWL-S API, the standard OWL-S precondition >> language, which will be "SWRLlike" as mentioned before. >> >> We'll have RDQL syntax support as well...so I guess you could use that >> if you really wanted to. > > Am I correct in assuming from your use of the future tense that this is > still in the pipeline? As OWL-S 1.1 is still in the pipeline, yes. The *capability* to perform conjunctive swrl-atom query against OWL-DL kbs is there in a hopeless inefficient form, in the current versions of Pellet. I believe you can use RDQL to invoke that. We are currently fully implementing the optimizations discussed in: http://www.mindswap.org/papers/SWS-ISWC04.pdf And it will be in a near term release. > If not, are there any examples available of how > to use Pellet to evaluate SWRL-like conditional expressions. Let's be careful. We don't (currently) evaluate SWRL *conditionals* (much). We evaluate SWRL like conjunctions (as queries). And we don't yet have the optimizations in our query compiler to make it really worth doing. However this will be forthcoming shortly. We welcome contributions, of course :) Ian Horrocks has some students working on this sort of thing as well. > I'd be very > grateful for any additional information about this. http://www.mindswap.org/papers/SWS-ISWC04.pdf >> I was thinking that a planner could do this. Of course, it wouldn't do >> precisely what you were asking for, since it doesn't determine >> compatibility in the general case of the *expressions*. But for the >> purposes of finding compatible *processes* for *composition* it seems >> sufficiently useful. >> > > The problem with planners is that compatibility of preconditions and > effects is based on (lexical) name matching. Well, unificiation with a KB. Or, in our case, conjunctive query against an OWL Ontology. [snip] > More generally, has there been any attempt to translate OWL-S > preconditions/effects to SHOP2 syntax? For example, to convert an > expression such as: [snip] See the above paper. We did this for a specific domain. As I said before, we'll have precondition compilers from a number of surface syntaxes as they settle. Comments, of course, welcome. Cheers, Bijan Parsia.
Received on Friday, 25 June 2004 13:35:28 UTC