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

RE: [UCR] RIF needs different reasoning methods --> PRs and assert

From: Vincent, Paul D <PaulVincent@fairisaac.com>
Date: Mon, 13 Mar 2006 00:17:19 -0800
Message-ID: <B3636F07C8359844A9A2370C5EA08CCBD07A8D@SRFMSGMB00.corp.fairisaac.com>
To: "Gerd Wagner" <wagnerg@tu-cottbus.de>, "Sandro Hawke" <sandro@w3.org>
Cc: <public-rif-wg@w3.org>

Thanks Gerd: interestingly, there are different approaches to PRs
regarding "asserting facts":

1. JESS directly supports fact assertion (probably from its OPS
background). See
http://www.jessrules.com/jess/docs/52/language.html#knowledge  

2. Blaze Advisor has NO direct support for "assert" or even simple
"fact" tuples: as a "commercial" PR engine it directly supports instead
the concept of a "business object model" based on an IT object model (eg
XML DOM, Java, .NET etc). So instead of 
    (assert (foo A B)))
We have
    set foo.A to B
... with the important caveat that previously I have defined the
association rule (via a business object model) (/terms and facts) that
foo is an object or class with attribute A with a type compatible with
B.

The significance here is that there is NO differentiation between assert
and retract: to retract a fact I simply re-assert a new value (or null).

In other words, commercial PR engines (I believe JRules is the same) do
not differentiate between assert and retract. (This may or may not be
significant in the context of this discussion...)

Paul Vincent
Fair Isaac Blaze Advisor --- Business Rule Management
OMG PRR and W3C RIF for rule standards
 -----Original Message-----
From: public-rif-wg-request@w3.org [mailto:public-rif-wg-request@w3.org]
On Behalf Of Gerd Wagner
Sent: Sunday, March 12, 2006 11:41 PM
To: 'Sandro Hawke'
Cc: public-rif-wg@w3.org
Subject: RE: [UCR] RIF needs different reasoning methods 


> What about production rules where the only action is assert?  
> Are those
> too different from Horn to count as "Roughly Horn" or are 
> they just not interesting enough to bother with.

They are too limited to be of an use. For instance,
while a virtual view (in a database) could be defined
with the help of a "Horn derivation rule", the
corresponding materialized view could not be defined
with a corresponding "Horn production rule" (because
keeping the materialzed view up-to-date would require 
a complementary retraction rule).

But this question should really be answered by our
production rule experts from ILOG and Fair Isaac.

<snip>
Received on Monday, 13 March 2006 08:18:11 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:27 GMT