Newsgroups: comp.robotics
Path: brunix!news.Brown.EDU!noc.near.net!howland.reston.ans.net!vixen.cso.uiuc.edu!cs.uiuc.edu!kadie
From: kadie@cs.uiuc.edu (Carl M Kadie)
Subject: Re: "That's not robotics; that's toy building"
Message-ID: <CBtBHI.38B@cs.uiuc.edu>
Organization: University of Illinois, Dept. of Comp. Sci., Urbana, IL
References: <CBHzGn.DI6@cs.uiuc.edu> <247cg1$9e4@wampyr.cc.uow.edu.au> <GERRY.93Aug10100832@onion.cmu.edu> <1993Aug12.190036.24847@phx.mcd.mot.com> <CBq6n8.IvL@cs.uiuc.edu>
Date: Sun, 15 Aug 1993 18:03:17 GMT
Lines: 70

Here are a robotics researcher's views on hobby robotics, posted here
anonymously. I received this via email. I'm posting excerpts with
permission of the author. 

(The author's views are not necessarily my own. So if you quote this,
please don't attribute it to me.)

================================================================
[...]
well, i have looked over at [comp.robotics], and it seems that this
toy vs. real robotics topic is causing quite a commotion. most of what
i read seems to me not to get at the key issues, i.e., those issues
that the so-called real roboticists use to make the evaluation of toy
vs. real.

it isn't that aaai robots are, in some sense, down-to-earth and
practical.  it doesn't have anything to do with purpose or intent
[...]

here's the issue...  aaai robots are, in their complexity, toys.  if a
high school student goes to radio shack, buys a few chips and a
circuit board and fabricates a basic computational circuit, e.g., a
simple processor, we don't say that he is doing computer science.  he
is a hobbyist.  he may learn something about soldering, but the truth
is that today's computer engineers don't do much soldering (they use
cad programs to do layout design, boards are fabricated in extravagant
ways --- soldering is simply no longer time effective in the computer
engineering world). he may learn something about circuit design and
switching algebra, but he will most likely learn less (and in a less
efficient manner) than he will learn in [class] when he arrives at
[the university] in a few years. in short, the high schooler has made
zero contribution to computer science, computer technology, or
computer programming.  he has learned very little of any value for his
future educational goals.  he has merely had a good time.

the analogy holds for robots.  we already know how to build toy
robots. in fact, we know how to build fairly large and complicated
robots.  the issue isn't where to solder, how to design digital
circuits, etc..  these are very low-level issues.  if a robotics
person needed to actually build his own control circuit (which
is unlikely, given the large number of computer interfaces available
for robots these days), he would use an off-the-shelf processor
board.  if a robotics person needs a robot arm, he buys an arm,
if he needs a mobile base, he buys one, etc...  the aaai robotics
competition, and hobby robotics, is just like digital electronic
hobbies ---  they are fun, but so  very simple by comparison to
the state-of-the-art that they have little to no value to the
robotics world, either pedagogically or technologically.   i agree
that it's good fun to build little robots from legos and discrete
digital components, but it isn't what robotics is about.

it isn't clear to me why a bunch of hobbiests even want to be called
'real roboticists.'  it also isn't clear to me why the ai community
hasn't learned the lesson of strips -- i.e. that building toy systems
while ignoring the many complexities of domains (in robotics, these
complexities are things like dynamics, control, world modeling,
sensing {not simple collision detection, but sensing forces, visual
sensing, etc..}, motion planning, etc.)  ultimately leads the
community to solve over-simplified problems that are of very little
value, and which leads to a devaluation of the discipline by the
larger scientific/engineering research community.  hence, [senior
researchers] of the field have to go around the country defending ai.

so, there's my input... as you can probably tell, i'm fairly annoyed
by the ai community's approach to robotics.  i'm happy for such
contests to be held, just annoyed by the hype and pretense.
=================================================
-- 
Carl Kadie -- I do not represent any organization; this is just me.
 = kadie@cs.uiuc.edu =
