Alt-BEAM Archive

Message #05084



To: beam@sgiblab.sgi.com
From: Bob Shannon bshannon@tiac.net
Date: Wed, 07 Jul 1999 19:22:26 -0400
Subject: [alt-beam] Re: Someting funny with the subject line? was Something funny with the


Steven Bolt wrote:

> On Wed, 7 Jul 1999 JVernonM@aol.com wrote:
>
> > Yes, I see your point. One could even say that Tilden himself
> > relies on the trial and error method to make these things.
>
> Afaik Mark T. tries very hard to do better than that.
>
> > Further, the bots themselves are engineered to use trial and
> > error to function randomly. It would seem that BEAM IS trial and
> > error. I'm not sure how you change that if it is so all
> > encompassing.
>
> BEAM `bots may be charming, remind us of living insects, be
> skillfully built or even art. But any behaviour, any trial & error
> we see is entirely between *our* ears. They `try' to move about and
> make `errors' as much as an aeolian harp tries to make music and
> errors the tune...

At what point does a good simulation start being a fair emulation?

> > And I really am having problems seeing how BEAM would ever come
> > up with, say, a manipulator that can pick up a certain object,
> > ignoring others and doing something meaningful with it using that
> > criteria.
>
> BEAM `bots don't have to perform tasks, but they might be more
> interesting when based on properly designed, described and
> explained circuitry. That kind of stuff can behave just as
> `insect-like', if that's what is wanted.

Technically, if it does not perform a tast, its not a robot. Make a
photovore
carry around a solar powered thingy, even a flag spinner, and now you have
a real robot.

But BEAM overlooks this a bit.

> > But, aside from all that, it is still a great way to pull kids
> > away from video games and get them to pick up a soldering iron.
> > Getting them to realize that they are doing weird things and that
> > there is a conventional method to these components is the hard
> > part.
>
> Which present BEAM makes harder than necessary.

That right. Circuit evolution can be seperate from robot evolution.
If SE's were cheap, bitty little things the size of a transistor, beginners

would go crazy building all sorts of things we would never dream of.

And thats probably why its never been done.

> > I do wonder if it really matters in the end. If a newbie
> > wishes to learn "proper use" they will. If they just want to
> > build something cool without knowing what's going on inside, they
> > will. Perhaps, just getting them interested is the real hurtle.

I totally agree.

> > And BEAM puts them over that one quite well. There are many
> > online who I have seen start out with asking "which way does a
> > resistor go in" who are now discussing advanced neural nets. Not
> > many, but they are there.
>
> It seems to me that fewer would turn away in frustration, and more
> would learn to really enjoy mechatronics if BEAM would put them not
> so much in the dark to begin with.

Then make the parts avaialble so that people can deal with mechatronics
rather
than how to free form an SE.


------------------------------------------------------------------------

eGroups.com home: http://www.egroups.com/group/alt-beam
http://www.egroups.com - Simplifying group communications



Home