From: Harold Boley (boley@informatik.uni-kl.de)
Date: 11/12/03
Hi Benjamin, > o variable naming: string/literal vs. a local URI > > currently in RuleML syntax is a string/literal, > we can keep that in the short term for the XML version > > disadvantage in RDF view: don't want to declare a literal > to be a variable for the entire Web universe > > local URI is local to a document, e.g., rulebase > > Harold: can view it as defining a separate alphabet of variable names, > local to a document > > this requires a change to the working draft of the OWL RuleML document The XML version's <swrlx:var>x1</swrlx:var> etc. was fine. The RDF version's <owlr:argument1 rdf:resource="#x1"/>, referring to <owlr:Variable rdf:ID="x1"/>, was accepted for the time being, e.g., because of the below-mentioned possibility for confusion. "Local URIs" used for implicitly universally quantified rule variables should actually be local to every single rule of a rulebase, to reflect the (narrow) scope of logic variables. Today's RuleML SC telecon will also talk about this. BTW, I updated Section 6. Mapping to RDF Graphs (http://www.daml.org/rules/proposal/rdfsyntax.html), and now hand it back to Ian, Peter, and Mike. TBD: We should discuss the correspondence between the XML and RDF syntax. We should explain that, e.g., <owlr:Variable rdf:ID="x1"/> is not to be confused with a "global variable declaration" in programming languages, where types and initializations could be added: it just declares that "x1 is in the alphabet of variables". We could explain why the other ('literal') option was not chosen, at least for the time being. We could show an actual RDF *Graph*. We could try to show how *any* rule can be mapped this way. Best, Harold
This archive was generated by hypermail 2.1.4 : 11/12/03 EST