W3C home > Mailing lists > Public > public-rif-wg@w3.org > June 2008

Re: ISSUE-65 (FINAL): What halting test should PRD cover? [PRD ]

From: Gary Hallmark <gary.hallmark@oracle.com>
Date: Mon, 23 Jun 2008 22:32:01 -0700
Message-ID: <486086D1.9080109@oracle.com>
To: Rule Interchange Format Working Group WG <public-rif-wg@w3.org>

systems commonly provide a way to halt after a maximum number of 
inference cycles have been performed

Jess actually has a mode called "runUntilHalt" where it does not stop 
unless a special rule action is executed.  This is used for control 
applications, e.g. a power plant cooling system where temperature 
"facts" may arrive at any time.  Rather embarassing to halt too soon here...

Rule Interchange Format Working Group Issue Tracker wrote:
> ISSUE-65 (FINAL): What halting test should PRD cover? [PRD ]
>
> http://www.w3.org/2005/rules/wg/track/issues/
>
> Raised by: Christian de Sainte Marie
> On product: PRD 
>
> Due to the action part in the rules, and the Retract in particular, the semantics of production rule systems that PRD covers does not guarantee that the execution of an arbitrary ruleset halts by starvation (that is, by the absence of further rule instances to fire). OMG PRR does not specify any halting test, only mentioning in the description of the semantics that the cycle "is repeated until some state is met".
> - Should starvation be the only halting test covered by PRD (in which case the question of halting is pushed to ISSUE-63: PICK) or should other halting test be covered as well?
> - If not only starvation: what halting test should be covered? What combination?
> - How should the intended halting test be notified to a RIF consumer?
>
>
>
>   
Received on Tuesday, 24 June 2008 05:35:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:49 GMT