[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: Shortcomings of Predicate Logic? (was RE: RDF

• To: 'Paul Prescod' <paul@p...>, xml-dev@l...
• Subject: RE: Shortcomings of Predicate Logic? (was RE: RDF Interpretation of XML documents )
• From: "Bullard, Claude L (Len)" <clbullar@i...>
• Date: Tue, 20 Aug 2002 16:47:51 -0500
• Cc: tbray@t...

All true, Paul.  And I would expect a rules-based 
processor to outrun a relational system without 
a rules base.  But put one on top of one, and it 
will quickly outperform a rule system sitting 
atop a stack of HTML documents.  The data model counts.

But it is exciting to teach a computer to compute 
in such a way as that it can exhibit behaviors 
that are much like what one observes as a result 
of a human thinking.  Again, do submarines swim?

This is not semantic web stuff.  But some technologies 
that the semantic web initiative is producing may 
be useful to the effort of creating semiotes.  The 
question then is, what would one use semiotes for? 
(credit:  the term "semiote" was introduced by 
Sylvia Candelaria de Ram).

Artificial personalities.  Companions.  It doesn't 
matter that they don't think as we may think.  They
are comforting, sometimes wise, and occasionally 
very funny.  What else do you need from a robot dog? 
Ok, it should guard the house too, so it need to 
know when to be friendly and to whom and who to 
attack and how.

Waaaay out?  Not as far as I used to think it 
was.  But being human, "my logic is unreliable 
in these matters" as Sarek said. ;-)

len

-----Original Message-----
From: Paul Prescod [mailto:paul@p...]

"Bullard, Claude L (Len)" wrote:
> 
> ...
> 
> If the thing being modeled is human, we don't
> model a machine.  His reply cuts both ways.

It is unfortunate that the use of words like "meaning", "semantic" and
"infer" encourage people to think that either predicate logic or the Web
is about teaching computers to think like people. The semantic web
technologies will allow computers to draw very basic conclusions based
on carefully constructed inputs, just as Excel can draw very charts
based on carefully chosen numeric inputs. It is, in my opinion,
off-topic to rejoin: "But humans don't do it that way." Humans don't add
numbers in the way computers do either. So what? As long as they add the
numbers or make the inference they can help us to solve problems. If the
semantic web is roughly as intelligent as a relational database it will
have succeeded...and it seems to have already exceeded that point.

• Prev by Date: Re: Shortcomings of Predicate Logic? (was RE: RDF Interpretation of XML documents )
• Next by Date: namespace reprise
• Previous by thread: RE: Architectural Forms (was Re: XHTML 2.0 and the death)
• Next by thread: namespace reprise
• Index(es):
  • Date
  • Thread