AI project. =)
Jecel Assumpcao Jr
jecel at merlintec.com
Wed Aug 14 17:50:38 UTC 2002
Alan Grimes wrote:
> It may be too slow for a practical AI but it should still be possible
> to VALIDATE my ideas with what is available.
In the same sense that you might have "validated" Squeak using a 256
byte Altair computer, yes.
On Tuesday 13 August 2002 23:59, Joshua 'Schwa' Gargus wrote:
> Ah, I see. That is a reasonable reading of that passage. However,
> in his rebuttals of some criticisms of the essay
> (www.edge.org/discourse/jaron_answer.html), he cites several
> historical trends that were locally exponential, but did not continue
> (eg: work becomes obsolete by the year 2000). I don't think that he
> is asserting the existence of an unavoidable law that causes software
> to bloat; note he begins with the phrase "If anything...". He is
> using a rhetorical device to rebut Belief #5 (That qualitative as
> well as quantitative aspects of information systems will be
> accelerated by Moore's law).
I think he was trying to dash the hopes of some of his critics that
Moore's law will last forever.
> > but only that they are the reflection of the Von Neumann
> > architecture (where every single transistor and machine cycle
> > "counts") on our current programming style and languages. I know he
> > tried hard to create a different style and failed.
> Did he? I am unfamiliar with his contributions outside of the area
> of VR.
His language efforts weren't outside the area of VR:
There is a great interview with Jaron in the book "Programmers at Work"
(Susan Lammers, Microsoft Press, ISBN 0-914845-71-3) in which he talks
about his language, though he was rather secretive at the time. He even
mentions the Dynabook (so this isn't 100% off topic ;-)
This paper on teleimmersion (really, really off topic now but worth
reading) has some interesting paragraphs about how VPL came to be:
More information about the Squeak-dev