Syntactic issues breakout

Sandro: Consider the naming, identifying URI question. Name predicates, variables, 
constants, datatypes.

Mike Dean: local names should be considered.

Hassan: the exact form is the main question. Variables are encoded within the RIF 
namespace. What form it should take.

Alex: predicates vs. beans in production languages.

Sandro: age property of a Person. 

Mike Dean: URN for Java packages would be nice to have.

A binding mechanism could be considered. 

Age corresponds to URI. 

Hassan: A local vocabulary. In the RIF what these local constructs correspond to.

Sandro: two rulesets about houses. Mapping the entities here allows merging them.

Hassan: IDL, binding is required foo means this. Signature mapping is what we are 
searching for.

Dave: RIF talks house A calls house, B calls it home. 

Harold: individual constant, local name and URI. Provide additional data.

Axel: Normal RDF treatment of namespaces, does not see much problem here. Why 
should it be different?

Sandro: mapping of namespaces is it the RIF question.

Jeff: RDF a set of triples what do we call it.

Local URI’s with Global namespaces. What is the problem.

Jeff: RDF b-nodes represented by local id’s.

Sandro: Do many rule systems use namespaces declaration?

Mike: A mapping could be optional and that will become omitted and that could present a 
problem.

System->RIF. Modules should correspond to namespaces.

Hassan: It could be a local namespace. Modules hide local names.

Alex: There is conflict of interest: modules hide stuff, global URI open it.

Dave: Precisely, the hiding is not important.

Dave: No need to standardise the mapping. Mapping is not our business. There has to be 
a vocabulary of local libraries. Allow people to use metadata to carry that data. We may 
how. Second issue: how map data structures to rules. XML Schema, URI for age 
corresponds to URI in XML Schema. Christian has this in mind to have data mapping in 
addition to rules mapping. 

Dave: many categories. Local variables do not become global. Local variables are they 
all required. Internationalization question arises.

Harold: URI vs. IRI.

Dave: URI ref is almost the same as IRI.

Harold: predicates/functions, actions.

Dave: using standard libraries. 

Harold: operations as well.

Alex: same thing.

Sandro: explore the politics on the using IRIs.

Hassan: preserve existing systems and make them exchangeable. C# and Java can have 
the same classes.

Sandro: the software will need to know about the namespace.

Hassan: # refers to the current fragment.

Hassan, Sandro: it is fine.

Sandro: issuing namespaces is a separate issue.

Sandro: URI crisis has been resolved. 

Harold: Pat Hayes IKL’s proposal, interchange of names. Postpone resolution from logic 
to ontology.

Hassan: RIFRAF is important—this is where ontologies will be derived.

Sandro: namespace=URI prefix. Curies—using qname-like syntax for IRI.

Hassan: A notion of identifier is superseded by that. 

Sandro: SUMMARY OF THE CONSENSUS SO FAR. RDF-like approach to URI’s. We 
do not see problem.

Hassan: Curies, the way to go both for local names.

Hassan: Explains about local names being invisible, behind quantifiers.

Variables for datatype instance and abstract instance.

Hassan: In AST, everything is explicit.

Jeff: Datatypes are RDF/S resource. Individuals are only OWL Thing. 

Jeff: Gerd Wagner’s concern about mapping F-Logic to RIF. That was Christian’s 
suggestion to discuss.

Hassan & Harold: slots are required.

Hassan: As Dave was saying roundtripping is important.

Hassan, Jeff: discussing extensions.

Sandro: get familiar with Gerd’s example.

Sandro: talk about roundtripping.

Mike Dean: should roundtripping be a requirement. To be discussed at the plenary. It 
should be discussed in what form it could on the issues list.

Hassan: libraries for things, like integers. Which datatypes are supported? What are the 
construction blocks. 

Mike Dean: library of variables, datatypes, a list of those is important to have at the end. 
Deliverable of RIFRAF.

Dave Reynolds: talked about RDF mapping (UC8). A scaffolding of RDF is used for 
assigning interchange metadata for rules. For UC8, the lowest granularity was one 
condition. An open question is about how far down to push RDF.

Hassan: Canonical syntax is AST-based. Human readable syntax is not unique. 

Sandro: The could be problems with RDF wraping.

Mike Dean: Hassan supports mixing with the right level of separation.

Mike: RDF does not have scope. Everything is flat.