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

RE: Semantic Web Rule Language (SWRL) 0.5 released

From: Boley, Harold <Harold.Boley@nrc-cnrc.gc.ca>
Date: Mon, 1 Dec 2003 20:36:59 -0500
Message-ID: <10C94843061E094A98C02EB77CFC3287026DF719@nrcmrdex1d.imsb.nrc.ca>
To: 'Drew McDermott' <drew.mcdermott@yale.edu>, www-rdf-rules@w3.org

Hi Drew,

You wrote:
I have some trouble understanding the RuleML web site.  There are so
many alternatives syntaxes.  Is one of them, or are some of them,
"official"?  (This is an embarrassing admission for me to make since
I'm listed on the website as a participant.)

The "official" version is always the newest release linked at
http://www.ruleml.org/#DTDs-Schemas.
The RuleML DTD Version 0.8 has been quite stable for some time.
A RuleML DTD+Schema Version 0.85 with 'roles' from OO will hopefully be
released by 19 Dec, solving a long-standing XML Schema inheritance problem
(http://www.ruleml.org/inxsd0.8.html#Issues).

You wrote:
But the argument list in Lisp 'defun' _is_ a quantifier; it
establishes a scope for variables as surely as 'lambda' does.

Yes, by (top-level) convention. Similarly, your modified Example 6.1-2 in
http://lists.w3.org/Archives/Public/www-rdf-rules/2003Nov/0178.html
(where your embedded <ruleml:Imp> should be called <ruleml:Implies>)
by (top-level) convention becomes the original Example 6.1-2 in
http://www.daml.org/rules/proposal/rdfsyntax.html
(since <ruleml:Imp> establishes a variable scope like <drs:Forall> does).

In RuleML 0.8 we allow <_rlab> (rule label) as a handle for an <imp>, e.g.
for prioritization (http://www.ruleml.org/dtd/0.8/ruleml-datalog.dtd) and
URI grounding (http://www.cs.unb.ca/~boley/ruleml/ruleml-rgs.pdf).
We do not yet have any embedded <implies>, but for those -- e.g. for
"implicational goals" in the body -- a label would not make much sense.

You wrote:
The free-variable convention has too many subtle "gotchas."  Even in a
well understood area like first-order theorem proving, I've seen
people get confused about why it is that queries have free variables
that are existentially quantified, or why backward chaining through a
rule like p(?x,?y) & p(?y,?z) -> p(?x, ?z) changes ?y from universal
to existential.

Well, that may be due to more operational than model-theoretic training.
RuleML since 2002-04-02 has had a (top-level) existential <query> element
(http://www.ruleml.org/#Queries).

But again, I think we should put a version of your (DRS) suggestion into
the SWRL (and RuleML, ...) issues lists.

Best,
Harold
Received on Monday, 1 December 2003 20:44:08 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:53:11 GMT