W3C home > Mailing lists > Public > www-rdf-rules@w3.org > November 2003

Re: Rules WG -- draft charter

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Mon, 17 Nov 2003 09:03:32 -0500 (EST)
Message-Id: <20031117.090332.01850278.pfps@research.bell-labs.com>
To: sandro@w3.org
Cc: pldms@mac.com, www-rdf-rules@w3.org

From: Sandro Hawke <sandro@w3.org>
Subject: Re: Rules WG -- draft charter 
Date: Wed, 12 Nov 2003 09:29:09 -0500

> 
> 
> > I guess my question is: is the rdf rule WG going to produce simply a
> > means to express rules, or something more like a rules programming
> > language (a la prolog, N3)?
> 
> Quite seriously: what's the difference?

Between which if the four above?

> And where there is a difference, which path is more important to the
> semantic web?
> 
> My answer: I think the language needs (unlike Prolog, but like N3) to
> be (1) monotonic and (2) have ordering of the rules be not
> significant, so that rule sets can be arbitrarily combined while
> maintaining their truth value.  

I believe that these two characteristics, the second in particular, answer
your question above.  My view is that ``rules'' have a declarative
semantics and ``rules programming languages'' do not, or, at least, that
the declarative semantics of rules programming languages is orders of
magnitude more comples than the declarative semantics of rules.  (Note that
forward chaining rules (without order, without negation-as-failure, without
cut, without ...) do have a declaractive semantics, in the form of a fairly
simple proof theory.  Yes, proof theories can be declarative, I just happen
to think that they are usually inferior to model-theoretic semantics.)

> There may be places where these
> requirements can be waived (especially for negation-as-failure
> (prolog's normal "not"), but they'll have to be circumscribed (in the
> general not necessarily technical sense of that word).  How and when
> they can be waived should be (IMO) out of scope for this WG.  Phase 2.
> Maybe the WG should be told to keep a path open to that stage....

Hmm.  Yes, closure (circumscription, NAF, ...) can be very useful.  It
would be nice to be able to get most of what is wanted here, but without
paying the cost that comes along with these sorts of constructs.  

>      -- sandro

peter
Received on Monday, 17 November 2003 09:03:44 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 11:10:15 UTC