[Top] [All Lists]

Re: [ontolog-forum] [ontolog-invitation] Ontolog Invited Speaker: R V Gu

To: "Peter Yim" <peter.yim@xxxxxxxx>
Cc: ontolog-forum@xxxxxxxxxxxxxxxx, "[ontolog-invitation]" <ontolog-invitation@xxxxxxxxxxxxxxxx>
From: sowa@xxxxxxxxxxx
Date: Wed, 30 Nov 2011 18:05:41 -0500 (EST)
Message-id: <dbf03d384eb5b10a00eee12767445489.squirrel@xxxxxxxxxxxxxxxxxxxx>


I'm traveling this week, and I'll be in meetings on Thursday.  So I won't be able to call in for Guha's talk.

But I have two questions that I hope somebody would ask.

Guha had been the associate director at Cyc, and he felt that CycL  was too complicated for most people.  I certainly agree with that.

But his solution was to develop a very simple subset of  logic,  which people could learn to use.  Then he hoped that they would gradually be able to migrate to a richer subset of logic as time went on.

That simple subset was triples, which in LISP (the base language of Cyc) would indeed be simple:  (A B C).  Unfortunately, RDF became the most outlandishly complicated way of saying (A B C).

So far, the only successful extension beyond RDF is RDFS.  People  have been pushing OWL for years, but with limited success.  If you look at the OWL ontologies published on the WWW, 95% of them do not use any features that go beyond Aristotle's syllogisms.

Suppose the W3C had adopted the following logic in 1998 and expressed it in LISP notation:

1. Triples and n-tuples expressed as (R A B) or (R A B C D...).

2. Aristotle's syllogisms (expressed in controlled English) for expressing and reasoning about the ontology.

3. If-then rules for a Horn-clause logic similar to Datalog for reasoning about the triples and n-tuples.  It could serve as query language with deductive  capabilities that are a superset of SQL or SPARQL.

4. Bindings to SQL for storing and retrieving n-tuples.

This logic and its _expression_ in a LISP-like notation could be presented in a very readable 10-page manual.  It would be much easier  to teach, learn, and use than the current SemWeb offerings.  It could be used in conjunction with relational DBs, object-oriented DBs, or RDFa tags in web pages.  And it could support a natural growth path to more sophisticated logics, such as CycL and others.

Question #1:  Why didn't Guha specify that in 1998?

Question #2:  Why doesn't he do that today?


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    (01)

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