[CL] Proposals for XCL 2.0

[Boley, Harold]

John,

Thanks for your greatly inspiring talk at RuleML 2010, and for all
the helpful discussions with the RIF and RuleML communities there.

I'm now subscribed to this email list, but not sure about the others
in the Cc (so I kept them explicitly in this first email).

Here are some of the RuleML features we discussed, which will help
to use it for CL serialization (cf. http://ruleml.org/1.0/glossary):
* Polyadic predicates and functions (allowing 'rest' variables)
* Typed variables (or unary predicates with attribute sortal="yes")
* FOL expressiveness (cf. http://www.w3.org/Submission/FOL-RuleML)
* Higher-order syntax (e.g., predicate variables)
* Reify (cf. http://www.w3.org/Submission/SWSF-SWSL/#ruleml-reification)
* Modal extension (preliminary, but will be required by SBVR)

Also, we started to develop the two CL RuleML examples below.

Looking forward to this cooperation.

Harold




CLCE:

Every cat is on a mat.


CGIF:

[Cat: @every*x] [Mat: *y] (On ?x ?y)


RuleML (fact with two typed variables):

<Forall>
  <Var type="Cat">x</Var>
  <Exists>
    <Var type="Mat">y</Var>
    <Atom>
      <Rel>on</Rel>
      <Var>x</Var>
      <Var>y</Var>
    </Atom>
  <Exists>
</Forall>


RuleML (rule with two sortal monadic predicates):

<Forall>
  <Var>x</Var>
  <Implies>
    <Atom>
      <Rel sortal="yes">Cat</Rel>
      <Var>x</Var>
    </Atom>
    <Exists>
      <Var>y</Var>
      <And>
        <Atom>
          <Rel sortal="yes">Mat</Rel>
          <Var>y</Var>
        </Atom>
        <Atom>
          <Rel>on</Rel>
          <Var>x</Var>
          <Var>y</Var>
        </Atom>
      </And>
    </Exists>
  </Implies>
</Forall>




CLIF:

(exists
  ((x Drive) (y Chevy) (z Old))
  (and (Person Bob) (City "St. Louis") (Agnt x Bob)
       (Dest x "St. Louis") (Thme x y) (Poss Bob y)
       (Attr y z)
  )
)


RuleML (Person, City sortal="no" by default, St. Louis needs no quotes):

<Exists>
  <Var type="Drive">x</Var>
  <Var type="Chevy">y</Var>
  <Var type="Old">z</Var>
  <And>
    <Atom> <Rel>Person</Rel> <Ind>Bob</Ind>       </Atom>
    <Atom> <Rel>City</Rel>   <Ind>St. Louis</Ind> </Atom>
    <Atom> <Rel>Agnt</Rel>   <Var>x</Var>   <Ind>Bob</Ind>       </Atom>
    <Atom> <Rel>Dest</Rel>   <Var>x</Var>   <Ind>St. Louis</Ind> </Atom>
    <Atom> <Rel>Thme</Rel>   <Var>x</Var>   <Var>y</Var>         </Atom>
    <Atom> <Rel>Poss</Rel>   <Ind>Bob</Ind> <Var>y</Var>         </Atom>
    <Atom> <Rel>Attr</Rel>   <Var>y</Var>   <Var>z</Var>         </Atom>
  </And>
</Exists>




-----Original Message-----
From: John F. Sowa [mailto:sowa at bestweb.net]
Sent: October 24, 2010 2:08 PM
To: Discussion of ISO Common Logic Standard (ISO/IEC 24707)
Cc: Boley, Harold; Adrian Paschke; Antonino Rotolo; Benjamin Grosof; Donald Chapin; Said Tabet
Subject: Proposals for XCL 2.0

At the Semantic Technology Conference in June, I had a brief discussion
about Common Logic and RuleML with Harold Boley and Adrian Paschke.

I won't go into the details, but the result was that Nino Rotolo,
the organizer of the RuleML 2010 Symposium, invited Pat Hayes,
Chris Menzel, and me to participate in the Symposium.  The goal was
to discuss the possible use of Common Logic to define the semantic
foundations for RuleML and related rule systems.

The Symposium was held from October 21 to 23.  Pat and Chris weren't
able to attend, but I gave a talk about Common Logic, issues related
to RuleML, SBVR, and RIF, and issues about the XCL dialect of CL.
Following is the complete program for RuleML 2010:

http://www.csw.inf.fu-berlin.de/ruleml2010/tl_files/ruleml2010/program/RuleML-2010-Programme.pdf

And following are the slides I presented on Friday, Oct. 22:

    http://www.jfsowa.com/talks/clruleml.pdf

At the end of this note is a copy of my last slide, which invited
RuleML, RIF, and/or SBVR experts to participate in the work on
revising XCL.  After the talk, I spoke with various people at the
conference whose responses ranged from positive to enthusiastic
to actually volunteering to work on an XCL 2.0 specification.
Among them are people on the cc list above.

The longest discussion I had was with Harold and Adrian.  We talked
about options for an XML-based notation that would (a) express the
full CL semantics and (b) include a Horn-clause subset that could be
easily mapped by XML tools to and from RuleML and RIF.  Harold said
that he could produce a first-pass draft within a week.

There were also many concerns about a wide range of features
that are not included in current CL standard.  SBVR, for example,
includes modality.  The titles of various talks in the RuleML
program mention uncertainty, fuzziness, and defeasible reasoning.

I summarized a "wish list" of features in slide 15 of my talk.
In slides 16 to 23, I discussed the IKRIS project, the extension
of CL to IKL, and the use of the metalanguage option in IKL to
support the modal features of SBVR.

My recommendation is to adopt IKL for CL 2.0, but *not* to include
the wish list in the standard.  Lenat & Co. have said that IKL is
sufficient to support all of CycL, and they have found means to
support essentially everything in that wish list.  Therefore, SBVR
and other rule systems that require any of those features could
adopt similar methods of using IKL to support them.

John
___________________________________________________________________

Updating the CL Standard

Some minor updates, revisions, and corrections have been suggested
for the ISO/IEC standard 24707 for Common Logic.

One major proposal is to adopt the IKL extension for CL 2.0.

Another proposal is to replace the current XCL, which has had very
little use, with a notation based on the logic subset of MathML.

The Horn-clause subset of XCL 2.0 should be compatible with the
notations for RIF and RuleML.

Any experts in RIF and RuleML would be welcome to collaborate in the
design of XCL 2.0. See

    http://www.common-logic.org

Begin by subscribing to the email list and take part in the discussion.