Re: [RIF] [UCR]: What is the RIF (revisited)

From: "Ginsberg, Allen" <AGINSBERG@imc.mitre.org>
Subject: [RIF] [UCR]: What is the RIF (revisited)
Date: Tue, 7 Feb 2006 10:55:41 -0500

> Dear RIF-WGers,

[...]

> Here are some questions to think about:
> 
> 1) If the RIF is mainly (or just) a rule-language, how does it enable
> interchange, anymore than some other rule-language?
> 
> 2) Does the RIF need to be or include one or more rule-languages in
> order to be a RIF?
> 
> 3) If the RIF has the ability to specify or describe rules and
> derivation procedures belonging to one or more  rule-languages, does
> that mean that the RIF must itself "be" one or more fully-fledged
> rule-languages?
> 
> 
> Allen

All this fundamentally depends on what you think a rule language is.

My view is that a rule language is no different from any other formal
representation language.  (An opposing view would be that a rule language must
have as well a fully-worked-out procedural meaning.)  

Under my view, how could a RIF *not* be a rule language or family of rule
languages?  The RIF is going to have a formal syntax - otherwise how can it be
used to interchange rules (because there will be no way to parse the RIF)?  The
RIF is also going to have a formal semantics - otherwise how can it be used to
interchange rules (because there will be no way to determine what a particular
bit of the RIF means)?  

Now it may be the case that there is no single formal semantics for all the
constructs of the RIF.  This would be a shame in my view, but I suppose that
it may turn out that this is the best that can be done in Phase 2.

Peter F. Patel-Schneider

Received on Tuesday, 7 February 2006 16:58:47 UTC