RE: Reference vs import <-- RIF Core shortened

I agree with Paul: PRD should focus on rule interchange, and not on
class descriptions.

Is it true that there are several formats to describe classes on the
shelf? I don't know them very well, but at least there is UML. Doesn't
it use an XML format to describe the classes? We should first look at
coupling with existing formats.

With PRD, I rather understand that a receiver of a PRD document
(containing rules) knows how to interpret the names in the rules. It has
a mechanism to resolve the meaning of the names. The rule document is
weakly typed (the names have o type or domain information), the receiver
is responsible for that.

What do you think?

Changhai

-----Original Message-----
From: public-rif-wg-request@w3.org [mailto:public-rif-wg-request@w3.org]
On Behalf Of Paul Vincent
Sent: Sunday, November 23, 2008 3:42 PM
To: Gary Hallmark
Cc: Christian de Sainte Marie; Dave Reynolds; kifer@cs.sunysb.edu;
Patrick Albert; Boley, Harold; Adrian Paschke; Axel Polleres; RIF WG
Subject: RE: Reference vs import <-- RIF Core shortened


Hi Gary: what you say makes perfect sense. However, your use case,
translating from an RDF model (presumably with an RDFS interpretation)
into a Java object model, makes explicit the idea that RIF = rule +
term/fact interchange.

This somewhat extends the scope of RIF from "rule interchange based on
some assumptions about a data / object model" to "rule and data / object
model interchange".

My concern is really that this could be a major undertaking. My prior
assumption was that, for PRD anyway, we would initially consider
something like XSL/XML+RIF interchange, and maybe extend to other
object/data mechanisms, and possibly between such mechanisms, in future
versions, as required. 

Cheers

Paul Vincent
TIBCO | Business Optimization | Business Rules & CEP
 
> -----Original Message-----
> From: Gary Hallmark [mailto:gary.hallmark@oracle.com]
> Sent: 23 November 2008 05:58
> To: Paul Vincent
> Cc: Christian de Sainte Marie; Dave Reynolds; kifer@cs.sunysb.edu;
Patrick
> Albert; Boley, Harold; Adrian Paschke; Axel Polleres; RIF WG
> Subject: Re: Reference vs import <-- RIF Core shortened
> 
> Paul,
> 
> RIF actually has an object/data model.  It's not very rich, because
some
> WG members believed (erroneously) that they could specify RIF without
> *any* data model.  But of course we have Herbrand terms (flat
relations
> in Core/PRD) and we have objects with slots and classes.  And we have
> the XML schema datatypes.  That sure sounds like an object/data model,
> and RIF translators are obliged to translate between this RIF data
model
> and the target rule language data model.
> 
> Now, if we have an "import foo.xml" addition to RIF, then the
> translator's additional job is to "hook up" the facts in that xml
> document with the target rule language data model using whatever means
> it can (e.g. I like JAXB), but the ensemble must behave *as if* the
xml
> syntax was mapped to RIF syntax in a manner that we specify.
> 
> Paul Vincent wrote:
> > Gary - are you envisaging that RIF translators will also do
object/data
> > model translation?
> >
> > I have to say I didn't imagine that would be in scope - ie I assumed
> > that RIF would assume that the translation of any external
> > object/data/fact model would be handled by other means.
> >
> > Paul Vincent
> > TIBCO | Business Optimization | Business Rules & CEP
> >
> >
> >> -----Original Message-----
> >> From: Gary Hallmark [mailto:gary.hallmark@oracle.com]
> >>
> >
> >
> >> challenge accepted, so below
> >>
> >> Christian de Sainte Marie wrote:
> >> Here's a production rule I'd very much like to write if I'm trying
to
> >> translate between RDF and Java objects:
> >>
> >> if ?o # ?c1 and ?o # ?c2 and not(?c1 = ?c2 or exists( ?c ?o # ?c
and
> >>
> > ?c
> >
> >> ## ?c1 and ?c ## ?c2))
> >> then ConstraintViolation("found an object that cannot have a Java
> >>
> > Object
> >
> >> Model")
> >>
> >>>

Received on Monday, 24 November 2008 10:10:00 UTC