W3C home > Mailing lists > Public > public-rif-wg@w3.org > December 2007

RE: Reminder: pending discussion "membership" (pending discussion on ACTION-350)

From: Paul Vincent <pvincent@tibco.com>
Date: Fri, 7 Dec 2007 01:53:38 -0800
Message-ID: <8F4A4531BB49A74387A7C99C7D0B0E0503655285@NA-PA-VBE02.na.tibco.com>
To: "Dave Reynolds" <der@hplb.hpl.hp.com>
Cc: "Public-Rif-Wg \(E-mail\)" <public-rif-wg@w3.org>

> You took my quote out of context. 

Oops. Apologies. " In the case of the Semantic Web stack, of which RIF
is a part ..." I read as an unconditional declaration of RIF being a
part of the semantic web, which could be misinterpreted, and is
therefore a statement to avoid IMHO. 

> My point was there are different
> stacks RIF might want to work with each of which already has their own
> way of handling these things.

Fully agree. 

> Sounds like you have a use case to share! Christian has previously
said
> this would be very unusual in PR applications.

It may certainly be unusual for *interchangeable* PR applications, but a
subclass mechanism may be used as a convenient way to store a local
result in objects that are not needed outside the rule system.

Example: for electronic contract exchange, I may used a corporate
contract XML schema, but find that I want to store intermediate rule
processing results inside the rule engine for other rules to use. So I
may subclass Econtract to EcontractForServices and add an attribute
AppropriateSkillLevelApplied, computed from some other data. I use this
in the rules but the external class / XML schema is not under my
control.
For interchange I would probably solve this some other way eg with a
local XML schema of extensions, or I would force a fix to the external
schema, etc.

Most PR engines are implemented in a 3GL / import XML (etc) into a Java
representation, which allows them some flexibility over object model
definitions. So subclassing an XML-derived class is not a big issue. But
it is NOT of interest to try and standardize these mechanisms (I
suggest) in RIF. Hence I fully concur with Christian (PRD should not
bother with such a mechanism).

However, I can fully understand why an AI-type / knowledgebase
application would want to include / embed schema info into its
knowledgebase. Its just that this is not "core" to "RIF" IMHO.

Paul Vincent
TIBCO | ETG/Business Rules 
 

> -----Original Message-----
> From: Dave Reynolds [mailto:der@hplb.hpl.hp.com]
> Sent: 07 December 2007 09:22
> To: Paul Vincent
> Cc: Public-Rif-Wg (E-mail)
> Subject: Re: Reminder: pending discussion "membership" (pending
discussion
> on ACTION-350)
> 
> Paul Vincent wrote:
> >> Using your data modelling language of choice. In the case of the
> >> Semantic Web stack, of which RIF is a part, the answer is RDFS/OWL.
> >
> > Semantics schemantics. RIF includes non sem web
requirements/features
> > etc, so it at best is a "shared part" of the Sem Web stack. IMHO. :)
> 
> You took my quote out of context. My point was there are different
> stacks RIF might want to work with each of which already has their own
> way of handling these things.
> 
> At previous F2F we've agreed that the two important stacks it needs to
> work with are semweb and XML.
> 
> The sentence you quoted was saying "*if* you are working in semweb
then
> ...". Likewise the next one was saying "*if* you are working in XML
> Schema then ...".
> 
> >> I am convinced that including these primitives moves RIF from the
> > domain
> >> of rule interchange into that of data model interchange. Had that
been
> >> explicitly part of the RIF charter I am not certain we would have
> >> approved the formation of RIF.
> >
> > Funny I originally drafted a similar point in my other response, on
the
> > premise that Michael used the term "knowledgebase" instead of
rulebase,
> > and that a Knowledge Interchange Format should be considered an
> > extension to RIF, if a KIF is desired.
> >
> > But, pragmatically, quite often rulebases (eg in PR) include things
like
> > local variable definitions, and sometimes even local subclass
> > definitions, to simplify rule language expressions when the domain
> > schema is "fixed" and needs "extending" in the rule system. So I
have
> > some sympathy for Michael's position.
> 
> Sounds like you have a use case to share! Christian has previously
said
> this would be very unusual in PR applications.
> 
> > 1. For the RIF use cases, we would typically want (for PR) an XML
doc +
> > schema to form the factbase.
> 
> Exactly. So how are subclass relations supposed to be connected to XML
> Schema?
> 
> Dave
> --
> Hewlett-Packard Limited
> Registered Office: Cain Road, Bracknell, Berks RG12 1HN
> Registered No: 690597 England
> 
> 
> > 2. The RIF charter implies (per my reading) that anything outside of
the
> > definition of rule should be considered a non-core extension.
> >
> > 3. The AI / Sem Web community are divided on the value of class
> > membership constructs in BLD.
> >
> > My simple inference from this discussion that perhaps BLD should be
> > BLCore (no "schema features") and BLDialect (BLCore + "schema
> > features")?
> >
> > Paul Vincent
> > TIBCO | ETG/Business Rules
> >
> >
> >> -----Original Message-----
> >> From: public-rif-wg-request@w3.org
> > [mailto:public-rif-wg-request@w3.org]
> >> On Behalf Of Dave Reynolds
> >> Sent: 07 December 2007 08:52
> >> To: Michael Kifer
> >> Cc: axel@polleres.net; Public-Rif-Wg (E-mail)
> >> Subject: Re: Reminder: pending discussion "membership" (pending
> > discussion
> >> on ACTION-350)
> >>
> >>
> >> Michael Kifer wrote:
> >>> CSMA had an action to bug me about the ## feature :-)
> >>> I thought that others might also be interested, so I am including
my
> >>> arguments below.
> >>>
> >>> First, one needs to be able to specify that one class is a
subclass
> > of
> >>> another class **as part of the KB**.
> >> I disagree, at least if by KB you mean RIF rules rather than RIF
rules
> > +
> >> externally specified ontology or data model.
> >>
> >> Expressing data models or ontological models and any subClass
> > relations
> >> associated with them is not a RIF requirement.
> >>
> >>> For instance,
> >>>
> >>> student##person.
> >>> father(person)##person.
> >>>
> >>> In KB apps this is used for reasoning, not just as part of a data
> >>> model. How would one specify this info otherwise?
> >> Using your data modelling language of choice. In the case of the
> >> Semantic Web stack, of which RIF is a part, the answer is RDFS/OWL.
> >>
> >> In the case of XML Schema models then complex types can be related
by
> >> both extension and restriction in ways that don't neatly map to
> > subClass.
> >>> Here is a more sophisticated example: parametrised lists.
> >>>
> >>> list(?Subclass) ## list(?Super) :- ?Subclass ## ?Super.
> >>>
> >>> (List of FOOs is a subclass of lists of BARs if FOO is a subclass
of
> >>> BAR. We could have list(father(person)), for example.)
> >>>
> >>> RDF's subclassOf does not cut it because
> >>>
> >>> 1. It imposes additional axioms, which are not commonly accepted.
> >>> 2. It is also not even defined for classes specified using
function
> >> terms
> >>>    (like list(?Foo)).
> >>>
> >>> Both arguments are also applicable to the RDF membership
> > relationship.
> >>> I am convinced that throwing out these primitives serves no
purpose
> > and
> >>> will just gratuitously cripple the BLD.
> >> I am convinced that including these primitives moves RIF from the
> > domain
> >> of rule interchange into that of data model interchange. Had that
been
> >> explicitly part of the RIF charter I am not certain we would have
> >> approved the formation of RIF.
> >>
> >> Dave
> >> --
> >> Hewlett-Packard Limited
> >> Registered Office: Cain Road, Bracknell, Berks RG12 1HN
> >> Registered No: 690597 England
> >
Received on Friday, 7 December 2007 09:53:59 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:44 GMT