[Top] [All Lists]

Re: [ontolog-forum] OpenCyc OWL Files

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Thu, 19 Feb 2009 11:49:34 -0500
Message-id: <499D8D9E.7060201@xxxxxxxxxxx>
Ian, Mitch, and Pat,    (01)

I wanted to add some comments to Pat's example:    (02)

PH> [CycL] uses full FOL expressivity in a CL-like format that
 > looks like higher-order logic plus the context-logic-based
 > microtheories machinery. The Cyc reasoner also has several
 > hundred special-purpose reasoning modules, and the Cyc formalism
 > is sometimes adapted so as to take advantage of them.
 > For example, rather than writing the axiom
 >    (forall (x)(if (P x)(Q x)))
 > Cyc will prefer the form (genls P Q), which means exactly the
 > same but is recognized by the Cyc inference machinery and
 > treated in a highly efficient way.    (03)

Note that Common Logic can express rules that explicitly relate
those two representations:    (04)

    (iff (genls P Q)
         (forall (x)(if (P x)(Q x))) )    (05)

Rules like this can be used to relate one representation to another.
As Pat mentioned, Cyc has many different reasoning modules, which
are specialized for different subsets of logic.    (06)

Some people make claims that RDFS or OWL is much faster than Cyc.
But that point is either false or extremely misleading.    (07)

Out of those hundreds of reasoning modules, Cyc has some that are
limited to subsets very similar to RDFS or OWL.  On those subsets,
the Cyc reasoning modules are blindingly fast.  I don't have the
statistics, but for reasoning about those subsets, I would bet
on Cyc rather than the current technology for RDFS and OWL.    (08)

I believe that Cyc has adopted a very good strategy of *not*
forcing the knowledge engineers to adopt a restricted notation
for logic.  Instead, they use a very expressive notation and
use the metalevel reasoners to choose the appropriate reasoning
module(s).    (09)

I don't believe that Cyc is perfect, and there are many things
that I would do differently.  But I believe that anybody who
proposes a new system should explain why they aren't using Cyc.    (010)

John    (011)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (012)

<Prev in Thread] Current Thread [Next in Thread>