From: Benjamin Grosof ([email protected])
Date: 07/19/04
Hi Gerd and all, At 09:07 PM 7/15/2004 +0200, Wagner, G.R. wrote: >I don't understand why a rule label should be different from a URIref. >After all, a rule label is nothing else as the name of a rule, and in >the SW a rule should be considered a resource like a class or a >property, so it should be identified with the help of a URIref, >and it would be preferable to reuse "rdf:ID" as its identifier >attribute name instead of the "ruleml:_rlab" element name. > >Please let's be economic with introducing new names. We don't >need new identifier attribute names for each new SW language. >We can simply reuse "rdf:ID" and "rdf:resource". > >Btw, I don't think it makes sense to consider a rule name an >(OWL) individual. The rule is the individual, and its name >is the name of this individual, but why should we consider it >to be an indivual itself? It's useful to have the rule label be an individual (or more generally a term) in the language, so that rules can talk about its relative prioritization, membership in modules, provenance, etc. This is a prime reason why in RuleML the rule label is an ind (and in the courteous extension more generally an ind or cterm) child of the label, rather than just a URI href. This was discussed in detail at the time the design was made. I think this rationale still applies. Benjamin >-Gerd > > > -----Original Message----- > > From: [email protected] > > [mailto:[email protected]]On Behalf Of David Z. > > Hirtle > > Sent: Thursday, July 15, 2004 8:44 PM > > To: [email protected] > > Cc: [email protected]; [email protected] > > Subject: SWRL RuleML Suggestion > > > > > > Hello Mike and all, > > > > As was discussed in the last JC telecon, the RuleML XSD > > (http://www.daml.org/rules/proposal/ruleml.xsd) accompanying the > > current version of the SWRL proposal is not fully compatible with > > RuleML (http://www.ruleml.org/0.86). Namely, rule labels in > > the SWRL proposal require empty content with an href value: > > > > > > <xsd:element name="_rlab"> > > <xsd:complexType> > > <xsd:attribute name="href" type="ruleml:RuleName" > > use="required" /> > > </xsd:complexType> > > </xsd:element> > > > > <xsd:simpleType name="RuleName"> > > <xsd:restriction base="xsd:anyURI" /> > > </xsd:simpleType> > > > > > > or more abstractly (from Section 5): > > > > > > <ruleml:_rlab > > ruleml:href = xsd:anyURI (required) > > > > > Content: ( ) > > </ruleml:_rlab> > > > > > > This use of the href attribute is different from RuleML's > > "webizing" use. > > > > In RuleML the _rlab element does not allow an href attribute, > > but rather contains an ind(ividual). Allowing so-called "foreign" > > individuals is still preferable to shifting them from element > > content to > > attribute value. > > > > For _rlab's, it would be desirable to use an OWL instance (Individual) > > e.g. from http://www.daml.org/rules/proposal/swrlx.xsd, > > > > <xsd:element name="Ontology"> > > ... > > <xsd:complexType> > > ... > > <!-- Instances --> > > <xsd:element ref="owlx:Individual" /> > > ... > > </xsd:complexType> > > ... > > </xsd:element> > > > > together with webizing via the "name" attribute of OWL's Individual > > (from http://www.w3.org/TR/owl-xmlsyntax/#owls_Individual): > > > > <Individual > > name = xsd:anyURI > > > > > Content: (Annotation*, (type | DataPropertyValue | > > ObjectPropertyValue)* ) > > </Individual> > > > > Using these defintions, we can now use rule labels as the > > content of the > > _rlab > > element instead of as an href attribute value. The _rlab element > > declaration > > would simply be as follows in the proposal's ruleml.xsd: > > > > <xsd:element name="_rlab"> > > <xsd:complexType> > > <xsd:sequence> > > <xsd:element ref="owlx:Individual" /> > > </xsd:sequence> > > </xsd:complexType> > > </xsd:element> > > > > > > Thanks, > > > > David and Harold > > > > ________________________________________________________________________________________________ Prof. Benjamin Grosof Web Technologies for E-Commerce, Business Policies, E-Contracting, Rules, XML, Agents, Semantic Web Services MIT Sloan School of Management, Information Technology group http://ebusiness.mit.edu/bgrosof or http://www.mit.edu/~bgrosof
This archive was generated by hypermail 2.1.4 : 07/19/04 EST