Newsgroups: comp.lang.smalltalk,comp.lang.c++,comp.object,comp.client-server
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!oitnews.harvard.edu!purdue!lerc.nasa.gov!magnus.acs.ohio-state.edu!math.ohio-state.edu!uwm.edu!newsfeed.internetmci.com!EU.net!peer-news.britain.eu.net!sunsite.doc.ic.ac.uk!susx.ac.uk!news.bton.ac.uk!Simon.Shurville
From: sjs16@bton.ac (Simon John Shurville)
Subject: Re: Problem with mass-nouns in OOA/OOD/OOP
X-Nntp-Posting-Host: tomcat.itri.bton.ac.uk
Message-ID: <DJDH0B.9GE@bton.ac.uk>
X-Posted-From: InterNews 1.0@diamond.bton.ac.uk.
Lines: 42
Sender: -Not-Authenticated-[3206]
Organization: The Composer Project
References: <DJ7sMr.EJr@bton.ac.uk> <30C327AE.2DA1@toa.com> 
 <30AF370B.1AE@vmark.com>  <4a9p0q$47h@news.onramp.net>
Date: Sun, 10 Dec 1995 13:19:23 GMT
Xdisclaimer: No attempt was made to authenticate the sender's name.
Xref: glinda.oz.cs.cmu.edu comp.lang.smalltalk:31888 comp.lang.c++:164263 comp.object:42507 comp.client-server:14402

In article <4a9p0q$47h@news.onramp.net>
oa@iah.com (Rick Lutowski) writes:

> They have their work cut out for them (hopefully, _they_ have read
> Weizenbaum!)  Since they will likely be writing a lot of throw-away
> software, it is probably appropriate that they are not using OO.

Worse, I think that they are using a mixture of simulated evolution and
a high level programming language known as graduate student!
 
> >I suppose that what I really wonder about is what the limits to object
> >oriented knowledge representation are. At the moment OO is being touted
> >as one of a number of 'saviours' for AI and as an engineer I believe
> >this to be a good thing. But my most philosophical and academic side
> >still wants to question how far this paradigm can be pushed.
> 
> I suspect there is no difference between the limits of knowledge
> representation, and the limits of object-oriented knowledge
> representation.  That is, an object-oriented approach does not
> extend the limits of what can be represented in a digital computer,
> it is merely a way of structuring what can be represented in a way
> that has beneficial maintainability characteristics in certain classes
> of programming languages.  

And it can also be quite readable/comprehensable by humans. This goal
may be a chimera, though. Perhaps, as Proffessor Sloman used to say an
AI might need to be scruffy and maybe eveolutionary techniques which
deliver a mind that can not be deciphered by an external observer are
the only approaches that will deliver life-sized intelligence?
 
> It is possible (although not certain) that other "hardware" architectures
> might extend the limits of knowledge representation, such as a
> neural network architecture.  These other architectures can be modeled 
> in a digital computer but, again, it will be a model of less than
> perfect fidelity, and so other architectures may have a greater
> knowledge limit than current computer architectures.  The ultimate
> form of this question is the old "can computers think like people"
> question.  It is a question that is likely to be unanswerable
> until we figure out how exactly people think.  That will probably
> not occur in our lifetimes.  Therefore, I believe your question
> can only be speculated upon, and not answerwed, with our present
> level of understanding.

Amen! Fun to do though.

