W3C home > Mailing lists > Public > public-rif-wg@w3.org > January 2006

Re: RIF and QL

From: Francois Bry <bry@ifi.lmu.de>
Date: Fri, 27 Jan 2006 14:12:48 +0100
Message-ID: <43DA1C50.2090204@ifi.lmu.de>
To: W3C RIF WG <public-rif-wg@w3.org>

jos.deroo@agfa.com wrote:

>Well, maybe I'm sleeping, but when instead of your 2 rules (in N3)
>{?X a kb:customer. ?X a kb:paysCC} => {?X a cons:paying-customer}.
>{?X a kb:customer. ?X a kb:paysCash} => {?X a cons:paying-customer}.
>I use a single rule
>{?X a ?C. ?C owl:unionOf (kb:paysCash kb:paysCC)} => {?X a 
>then given the facts
>kb:customer rdfs:subClassOf [ owl:unionOf (kb:paysCash kb:paysCC)].
>:Paul a kb:customer.
>and given the rules
>I'm getting proof evidence (*) for
>:Paul a cons:paying-customer.
>What am I missing??
It seems to me that you are missing that Enrico's point was not about
finding a way to implement his example, but to cope with disjunctive
information in general. Enrico's goal requires deduction methods that go
beyond what many rule languages offer.

Many rule languages are based on construvctive logic, or if you like it
better, simple modus ponens (with forward or backeward chaining). For
coping with disjunctive information, one needs refutation, or techniques
like anscestor resultion amounting to refutation.

My understanding, is that Enrico's example is about the power of the
deduction methods used in a rule engine. RIF might well give rise to
express Enrico's example, copming along with a simple declarative
semantics (or model theory), without making any commitment/assumptions
about a procedural semantics.


PS: Enrico, you are right: I misunderstood at first your point. I apologize for this.
Received on Friday, 27 January 2006 13:12:58 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:47:36 UTC