Thanks, penni

[From Bill Powers (920601.0900)]

Penni Sibun (920531) --

The language of the "pengi" article was a bit strange to me -- I'll give it
a closer look. I don't recall seeing anything in it about behavior
controlling perception, but that concept may be hidden in it. I certainly
agree about making choices -- I don't think "decisions" play a very large
part in behavior except when we get into internal conflict.

Your precis of the main idea is interesting. We, of course, have been
fighting against the idea of behavior that you describe (plan then execute)
for a very long time, and control theory is almost entirely about systems
that live in a real-time world and interact with it as it changes. Is there
any concerted move in AI away from the old view?

Thanks for the updated addresses: I sent my post to the MIT address that
appeared on the paper. I'll resend it right now.

What's "imho?"

Best

Bill P.

[From Chris Malcolm]

I recommended Bill's "Behaviour as control of Perception" book to Brooks
and his group at MIT in 1987, and some of them later discussed it with
me and found it interesting. I can't remember if I ever discussed it
with Chapman, but I certainly have discussed it with Agre (who did his
PhD with Brooks) and briefly sojourned in Sussex. Nobody else, however,
can speak for Agre's opinion of anything :slight_smile: I mention it to students in
my robotics course here at Edinburgh (some of whom now post to this
group), and it features in current robotics research at Aberystwyth
University. I first heard about it from Karl Kempf in about 1982, lately
of Intel and not concerned with robots now, but once leading McDonnel
Douglas's assembly robotics research programme. He said it was popular
among roboticists who were interested in learning robot design
architectures from nature.

[From Chris Malcolm]

   I recommended Bill's "Behaviour as control of Perception" book to Brooks
   and his group at MIT in 1987, and some of them later discussed it with
   me and found it interesting. I can't remember if I ever discussed it
   with Chapman, but I certainly have discussed it with Agre (who did his
   PhD with Brooks) and briefly sojourned in Sussex. Nobody else, however,

no, chapman did his phd w/ brooks; agre's was w/ mike brady.

        --penni

[From Chris Malcolm]

penni wrote:

no, chapman did his phd w/ brooks; agre's was w/ mike brady.

You're quite right, in formal terms this is true, and Brooks was his
internal examiner (or maybe external once Brady moved from MIT), but at
least latterly, in terms of intellectual affinity, Agre was a great fan
of Brooks, they spent a lot of time together, and up to the time that
Brooks visited Brady at Oxford (partly for Agre's examination) Brady was
profoundly sceptical of Brooks's approach. After that visit Brady
softened to saying things like "the subsumption architecture has some
important deep ideas, but is very subtle and easily misunderstood",
whereas earlier he would say things like "Brooks has lost his marbles."