Re: RIF: Production and Event-Condition-Action Rules

From: Francois Bry <bry@ifi.lmu.de>
Subject: RIF: Production and Event-Condition-Action Rules 
Date: Fri, 02 Jun 2006 09:43:42 +0200

> A few thoughts on Production and Event-Condition-Action Rules and the RIF:
> 
> 1. Production and Event-Condition-Action Rules are a must for RIF
> because they are needed for many applications.

I believe that this is only necessarily the case under two extra
assumptions:
1/ That the widescale adoption goal must be met.
2/ That the widescale adoption goal cannot be met without having these
   applications. 
I do not believe that either of the assumptions are true.  I do not believe
that either of the assumptions are held by all the other members of the
working group.

> 2. Production and Event-Condition-Action Rules are imperative
> programming that use  "queries" or "conditions" as used in declarative
> or deductive rules. The semantics of Production and
> Event-Condition-Action Rules poses no problems as soon as the "queries"
> or "conditions" of declarative or deductive rule are given a semantics.

How so?  What happened to the "imperative programming" parts of these rule
formalisms.   Can't they cause problems?

> Admittedly, so-called "conflict resolution" (which instance of
> Production or Event-Condition-Action Rules to choose for firing when
> several might be selected) need to be specified.

Yes, indeed.

> 3. Production rules are well suited for non-distributed systems,
> Event-Condition-Action Rules are a must with distributed systems.
> "Events" give rise to exchange informations between eg Web sites.
> Event-Condition-Action Rules make it possible to specify at several Web
> sites a reactive behaviour depending on actions and event taking place
> at remote sites. Since the RIF is to be used on the Web, considering
> Production rules only, ie not considering Event-Condition-Action Rules,
> for the RIF,would be in my humble opinion a serious mistake.
> 
> Francois

Peter F. Patel-Schneider

Received on Friday, 2 June 2006 11:05:54 UTC