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

[RIF] Re: Work on JRules abstract grammar

From: Ginsberg, Allen <AGINSBERG@imc.mitre.org>
Date: Wed, 18 Jan 2006 10:26:25 -0500
Message-ID: <90A462F2D6E869478007CD2F65DE877C69D3C8@IMCSRV5.MITRE.ORG>
To: <public-rif-wg@w3.org>, "Christian de Sainte Marie" <csma@ilog.fr>, <Harold.Boley@nre-cnrc.gc.ca>
Cc: <Harold.Boley@nre-cnrc.gc.ca>

Hi Christian & Harold,

The point I was making just as we ended the call was that, TO START
WITH, I would definitely like to see one of the other rule vendors in
the RIF pick up on this idea, or maybe have someone volunteer to do
that for a "public" system like CLIPS (or JESS). (Also, I noticed that
neither of those rule-platforms appear on Christian's list
(http://www.w3.org/2005/rules/wg/wiki/List_of_Rule_Systems. Perhaps no
one in WG mentioned these in their use-cases, but that doesn't seem to
me to be a reason to exclude them from consideration.)

However, I don't think this should be limited to working on the
abstract syntax portion of the RIF.  Presumably these closely related
production-rule systems share a common meta-model of how a certain form
of rule-based inference mechanism works.  So besides just sharing an
abstract syntax, they should also share a common inference model. We
need to start thinking about how inference mechanisms are to be
expressed in the RIF.  Obviously, because these systems share a common
inference model, having it represented in a RIF meta-model is probably
overkill for the purpose of translating between THEM.  But when we
start thinking about interchange between such systems and
prolog-based-systems, i.e., systems whose inference mechanisms belong
to different meta-models, then the ability of the RIF to represent
inference mechanisms should be important in the development of
automated rule-interchange systems.  

Once we have got a meta-model that we are reasonably confident  covers
IRL and one or two of the others, then we should be able to come up
with test cases and specifications for a translator.

This should be the "easiest" case of rule-interchange enabled by the
RIF. 

Allen
Received on Wednesday, 18 January 2006 15:26:49 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:26 GMT