ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] owl2 and cycL/cycML

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Tue, 03 Aug 2010 13:12:22 -0400
Message-id: <4C584DF6.309@xxxxxxxxxxx>
Ian,    (01)

We have no disagreements about complexity theory.  That is a matter
for analysis and proof, not debate.  But the crucial issue is the
"knowledge acquisition bottleneck" which burst the AI bubble in
the 1980s.  New technology has been developed since then, and we
should ask how that technology can address the problems.    (02)

Following are slides on that topic, which I presented at a conference
last week:    (03)

    http://www.jfsowa.com/talks/futures.pdf
    Future directions for semantic systems    (04)

Some comments on your note:    (05)

 > What I say about robustness is nothing more than simple statement
 > of what follows directly from complexity theory.    (06)

I was questioning the word 'robust'.  People have complained that
all formal systems (including conventional programs) are notoriously
fragile.  Every major programming language is undecidable, but
programmers never try to reduce the expressive power.  Instead,
they develop better programming tools and methodologies.    (07)

 > You seem to say that it is best to allow users to write whatever
 > they like and then leave it to the reasoner to figure out how best
 > to deal with it.    (08)

My remarks were based on issues in the futures.pdf slides, which
address two major questions:    (09)

  1. What kinds of tools should we develop for the subject-matter
     experts -- people who are professionals, but not in AI or IT?    (010)

  2. How can automated tools enable "the reasoner to figure out
     how best to deal with" the knowledge base?    (011)

Question #1 is the major theme of those slides.  But Cyc and other
projects have done a great deal of work on point #2.  See slides 7
and 8 of futures.pdf.  See also p. 6 of the following paper:    (012)

    http://www.jfsowa.com/pubs/fflogic.htm
    Fads and fallacies about logic    (013)

Relational DBs, for example, automatically transform queries to
an optimal form.  Cyc does even more.  They have over 3 dozen
inference engines, and the system automatically chooses which
one(s) to use for any given problem.  Good algorithms can do
a much better job of "dealing with" the KB than the SMEs.    (014)

 > ... but you might equally well declare that only expert
 > programmers should be allowed to produce software.    (015)

As a matter of fact, the overwhelming amount of software in use
today was developed by expert programmers.  Students and amateurs
can and do write programs, but they are rarely used by anyone
other than the authors.    (016)

 > ... it is more realistic (and more useful) to try to develop
 > tools and methodologies that help people to do a better (if
 > not perfect) job.    (017)

I agree.    (018)

John    (019)

_________________________________________________________________
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    (020)

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