W3C home > Mailing lists > Public > public-rif-wg@w3.org > June 2006

RIF: can we all work together?

From: Vincent, Paul D <PaulVincent@fairisaac.com>
Date: Thu, 1 Jun 2006 05:59:19 -0500
Message-ID: <3FCAC95D3E99F54B98300F372AA1B2614B94FD@STPMSGMB01.corp.fairisaac.com>
To: "Gary Hallmark" <GARY.HALLMARK@ORACLE.COM>
Cc: <public-rif-wg@w3.org>

Gary - You would think that "rule interchange" would infer "working together" but possibly I am being na´ve! 

Comments below:

Paul Vincent
for Fair Isaac Blaze Advisor  -- Business Rule Management System
@ OMG and W3C standards for rules
> -----Original Message-----
> From: public-rif-wg-request@w3.org [mailto:public-rif-wg-request@w3.org]
> On Behalf Of Gary Hallmark
> Sent: Thursday, June 01, 2006 8:06 AM
> To: public-rif-wg@w3.org
> Subject: can we all work together?
> 
> 
> I'll just toss this out for constructive criticism--
> 
> Start with the Boley et. al. common condition syntax from
> http://lists.w3.org/Archives/Public/public-rif-wg/2006Apr/0068.html

[PV>] Can we assume "free variables" effectively map to PRR rule variables ie semantically map to Rete bindings. Also there is the issue of typed variables (typically required in PR used against object models).

In PRR you have the concept of scoped variables that are expanded at runtime (to bindings). For example: 

>From Harold's example in the link above:
Example 1 (A RIF condition in human-readable syntax):

  In this condition, ?Buyer is quantified existentially, while ?Seller
  and ?Author are free:

  And ( Exists ?Buyer (purchase(?Buyer ?Seller book(?Author LeRif) $49))
        ?Seller=?Author )

A mapping to a PRR is approximately (NOT using PRR OCL syntax to keep things easier to understand)

Predefined Types: the major difference over RDF is probably the Type issue. Types can of course be represented in RDF but this would be an implementation constraint...

Variables / Parameters to ruleset:
?Buyer is a variable.

Rulevariables and filters:
?Seller is any Person.
?Author is any Person.
?SpecifiedBook is any Book where author = ?Author and title = LeRif and price = $49
Conditions:
?SpecifiedBook.author = ?Seller
?Buyer.purchases contains ?SpecifiedBook

> 
> Add "heads" or "conclusions" or whatever you need to have logic-based or
> Horn rules and give these rules a model theory.  E.g.
> http://lists.w3.org/Archives/Public/public-rif-wg/2006May/0248.html
> These rules should be expressive enough to use in the following steps.

[PV>] I'm not sure I follow this step or why it exists. If you are proposing an arbitrary "head" for a PR (I would propose the PR rule name - indeed in the Blaze PR system the rulename is equivalent to a macro representing the rule conditions) then fair enough, but why?. This is meaningless from a PR usability perspective until you add the actions (step below). So what is the value of this step? Phase 1 "compliance"? Something to make the logicians more comfortable about PR?

> 
> Add production rule actions to the common condition syntax and give
> production rules a semantics using logic rules defined in the previous
> step.  See Hasan's email at http://lists.w3.org/Archives/Public/public-
> rif-wg/2006May/0115.html

[PV>] Yes, but won't this step break the logician's acceptance of the rules? (Not that that matters for RIF's wider goals of course, but I'm assuming the "actions" are what "breaks logical conformance" for logicians).
> 
> Continue for ECA rules, etc.
> 
> Cheers,
>    Gary
> 
> 

This email and any files transmitted with it are confidential, proprietary
and intended solely for the individual or entity to whom they are addressed.
If you have received this email in error please delete it immediately.

Received on Thursday, 1 June 2006 11:06:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:29 GMT