Alt-BEAM Archive
Message #00557
To: Steven Bolt sbolt@xs4all.nl
From: Bob Shannon bshannon@tiac.net
Date: Wed, 17 Feb 1999 20:05:16 -0800
Subject: [alt-beam] Re: hmmm...i wonder which subject
Steven Bolt wrote:
>
> On Tue, 16 Feb 1999, George Rix wrote:
>
> > That isn't impossible with BEAM, AFAIK. If you interface it with some sort
> > of control device, like a CPU, or a BASIC Stamp, and upload the program, it
> > should work (I don't know a thing about BASIC stamps, so don't jump on me if
> > you couldn't use it there).
>
> There has been quite a bit of talk about adding a programmable
> device to some combination of Microcores, Nervous Nets and so on to
> gain something, but I have yet to see a worked out example to prove
> the point.
Again, look at Terry Newton's work. He used a PIC to drive a BEAM smart
body.
AS it turns out, the microcore is not predictable enough for the CPU to
get
along with. Also the microcore does not do anything that cannot also be
done directly by the CPU, so why penalize the robot with the microcore
at all?
> Imho, once you go for a programmable, it's best to use uCs all the
> way. I recall Mark T. being asked about using Nervous Nets as an
> `autopilot' for cars; he answered that would not be a good idea,
> because Nervous Nets are not predictable enough.
> If you send a scout to Mars, you want it to its job predictably...
So whats the advantage of BEAM again?
------------------------------------------------------------------------
eGroup home: http://www.eGroups.com/list/alt-beam
Free Web-based e-mail groups by eGroups.com
Home