RE: Questions

[Alain ESCAFFRE wrote,]
> We are  french student fome the "Ecole Polythecnique de Nantes"
working
> on a project about a geometrical ontology. We aim at describing
> geometrical objects and axioms thanks to RuleML and OWL. We read the
> Draft Version 0.5 of 19 November 2003 about SWRL and this atempt to
> unify owl and ruleML for inference on  the semantic web is very
relevent
> and useful ! Here are some questions about this draft and SWRL :
> 
>     - Chapter 5, exemple 5.1-2 :
> <ruleml:imp> 
>   <ruleml:_body> 
>    <swrlx:individualPropertyAtom  swrlx:property="hasParent"> 
>      <ruleml:var>x1</ruleml:var>
>      <ruleml:var>x2</ruleml:var>
>    </swrlx:individualPropertyAtom> 
>    <swrlx:individualPropertyAtom  swrlx:property="hasSibling"> 
>      <ruleml:var>x2</ruleml:var>
>      <ruleml:var>x3</ruleml:var>
>    <swrlx:individualPropertyAtom  swrlx:property="hasSex"> 
>      <ruleml:var>x3</ruleml:var>
>      <owlx:Individual owlx:name="#male" />
>    </swrlx:individualPropertyAtom> 
>  </ruleml:_body> 
>...
> 
>    Does this property "hasParent"  refers directly to a property  of a
>owl ontology about family  or is just created here and right now
simply
>thanks to this line of code ?

"hasParent" atom in the example is refering to an ObjectProperty defined
in an OWL ontology. You have to define "hasParent" property explicitly
in an OWL document.

> if   we write code such like
> <swrlx:individualPropertyAtom  swrlx:property="belongTo"> 
> <ruleml:var>pointA</ruleml:var>
>       <ruleml:var>Line</ruleml:var>
> 
> Does it imply that  we have created this property  "belongTo" in the
> ontology we have  implemented before  in owl ?
 
Same as above. The individualPropertyAtom does not define a property.

>  -  About the issue list and and extensions, we read swrl doesn't
> manage negation, what about  constraints rules ? In fact our project
> (because of the axioms) needs both of of them so we are afraid not to
be
> able to use this Semantic Web Rule Language : ?
>

Support for negation in a semantic web rules language is a controversial
issue.
NAF is not accepted well by semantic web community for its unsafety.
Classical negation seems to be considered positively[1].

> 
> - Finally, in case you confirm me SWRL is only horn clauses oriented
for
> the moment (no negations no constraints programmation),  could you
give
> us a link where I find  a Rule ML exemple which  use owl objects ?
> because We have difficulties to make the connexion , difficulties that
> swrl would have erased !

In RuleML, you can refer to a web resource by "href" attribute, by which
you can use OWL classes or properties as RuleML predicate symbols.
You can also refer to OWL individuals as constants.

And there're saveral rule engine or reasoner implementations with which
you can refer to OWL resources when writing rules [2][3][4]. Some of
these support NAF.

Best,
Minsu

[1] http://www.daml.org/listarchive/joint-committee/1576.html
[2] Bossam: http://mknows.etri.re.kr/bossam
[3] Euler: http://www.agfa.com/w3c/euler/
[4] CWM: http://www.w3.org/2000/10/swap/doc/cwm.html

Received on Friday, 12 December 2003 00:35:56 UTC