COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\input memo.tex[let,jmc]
C00020 ENDMK
C⊗;
\input memo.tex[let,jmc]
\leftline{The Cult of Information: The Folklore of Computers and the True Art of Thinking}
\leftline{By Theodore Roszak}
\leftline{Pantheon Books, 238 pages, \$17.95.}
\vskip .2in
\leftline{By John McCarthy}
\vskip .2in
	It's natural that computers should be surrounded by hype ---
computers being the symbol of our information age.  Bah, I'm beginning
to sound like Roszak and the rest of them.  Start over.

	Computers are important, but there's a lot of exaggeration, and
Roszak has found plenty to complain about.  However, he
exaggerates the dangers and expresses his prejudices.
We Earthmen have learned to live with
advertisements, and while they sometimes succeed in getting people to
spend money in ways I regard as foolish, I never met anyone or
even read in the newspapers about anyone who really expects a promotion
from using the right soap.

	Here are some of the exaggerations Roszak found and some
of his own.  Here also is my estimate of the
phenomena he discusses and some of the issues he raises.

	By the way there are also a lot of ordinary inaccuracies, e.g.
about when IBM got into the computer business, how the ARPAnet got
started, what the phrase ``Information please'' meant, who are Roger
Schank, I.J. Good, Pamela McCorduck and Robert Jastrow,
when computer courses began in the universities, what the
initials DARPA stand for and whether the New York Times of January 13,
1985 has a page A1.  These errors aren't essential to Roszak's
arguments, but they arouse doubt.

\noindent The cult of information:

	Roszak cites assertions that this is the ``information
age'', that our economy is or soon will be based more on information
than on goods.  Implausible.  Ask yourself
what you would do with another \$20,000 to spend.  What
fraction would go for information and what for home improvements,
cars, travel, eating in restaurants, liquor, dope and supporting orphans
in the Third World.

	He quotes John Naisbitt's ``Megatrends'' saying ``The new power is
not money in the hands of the few, but information in the hands of the
many''.  Nonsense indeed, but it's hard to believe that people will buy
more information than they want.  Look at the collapse of the home
computer boom.

\noindent Personal computing:

	Roszak finds most of the proposed uses of personal computers
not worth the trouble and expense.  It's partly true.

	I have had a computer terminal in my office and at home connected
to the same computer since 1971.  I use it for all my writing, for electronic
mail and for
some programming --- I'm a computer scientist.  As for information services,
we have an Associated Press news wire coming into our computer, I have
used the Stanford University on-line catalog and various databases
available through dialog.  Most people probably wouldn't find this
combination of uses sufficient to justify buying a computer or maybe
even for maintaining enough expertise to keep using the services.
We tried paying bills by computer, but it didn't work too well.
Nice but not revolutionary; I'd give up my TV to keep it but not my car.
As long as only the present services are available, personal computing
will remain a hobby except for people who do a lot of writing, use
spreadsheets, or have some other professional reason to compute.

	However, I believe that most people will become computer users when
the full information resources of our society become available.  Here
are some examples.  To find out what stores carry a desired item at
any time of day or night by connecting to its on-line catalog.  To
be able to read any book or magazine in the Library of Congress, not just 
its catalog.  To send electronic mail to anyone in the world.
  To be able to comment on any article you read in a way
that permits any other reader of the article to summon up any comments
that have been made including yours.  It still may not be as revolutionary
as the car or telephone.

\noindent Computer literacy:

	Roszak doesn't think much of it, but this is entangled with
his idea that it will teach people to think in an algorithmic way
which will distract them from ``master ideas''.  His fears
are unwarranted.  Learning to think in one way doesn't prevent
people from thinking in other ways also, and he offers no evidence
that teaching computer literacy has had the effects he fears.

	Many of the proposals for computer literacy are indeed
intellectually deficient.  Some proponents say that computer literacy
involves overcoming people's fear of computers by letting them sit in
front of a terminal, log in, edit files and print papers.  This is useful,
but it doesn't amount to much intellectually or take much time.  The only
people I've met who had any real fear of using computers were executives
and other big shots.  They feared their underlings would get to laugh at
their mistakes.  The secretaries in our lab were using terminals for
word-processing in 1971 and none ever failed to master it, even
temporaries, some of whom had never heard of typing into a computer
before.

	Education in computers should emphasize intellectual
value, not immediate practical uses which are readily picked up.

	It must include learning to program.  It is an individual
intellectual breakthrough for each person to discover that it is possible
to decide what you want the computer to do and write it down sufficiently
precisely that this dumb machine will actually do it.  Just controlling
an editor isn't enough to really give the idea.  Another way of
looking at it is this: programming is how we humans talk to our mechanical
servants.

	What languages should we use?  If the teachers could cram it all
in, they should learn several languages.  An actual machine language
tells you what these beasts are really like.  A sequential language
like BASIC or FORTRAN tells you what most programming is really like.
LOGO or LISP tells you how to compute with
symbolic information and not just numbers.  Finally, logic programming
in Prolog lets you express at least some of what you want to tell the
computer in the form of facts.  Very smart high school students can
learn all this in a semester, but most can't.  Rather than drag it out,
less should be done.  If it has to be just one thing, I'd pick logic
programming as being closer to how we'll speak to the servants in the future.

 	While I'm at it I'll express the opinion that the de-emphasis
of Euclidean geometry in high school education in favor the pedantic
``new mathematics'' was a blunder.  There are no intellectual
surprises in new mathematics, and the student is rightly ungrateful
for being reminded of the obvious at tedious length.  I remember that
my high school geometry book had a proof of Pythagoras's theorem
attributed to President Garfield.  Here we have an individual whose
talents ran to law and politics, not mathematics, and yet the
intellectual interest of the subject drew him to finding a new,
and somewhat surprising proof, of a 2000 year old theorem.

	Roszak doesn't mention one important bad effect computers have had
on education, especially at the lower levels.  I think teachers aren't as
smart as they were when I went to school, and the reason is that too many
of the granddaughters of the smart women who taught me are computer
programmers.  They make better money, and the work isn't as demanding.
Something can be done to make entry into teaching easier, e.g.  by
eliminating required courses in education, but the great expansion of
opportunities for intellectual work means that we must learn to do more
education with fewer highly qualified people.

\noindent Computing and the ``true art of thinking'':

	I'm the computer scientist Roszak cites who said it will take
from five to 500 years to reach human level artificial intelligence.
There are fundamental conceptual problems to be solved that haven't
been properly identified yet.  Maybe some clerk in the Swiss patent
office has just solved them and hasn't yet published his paper.  Maybe
they won't be solved for centuries.  There is no reason to suppose
that reaching human level artificial intelligence is an easier problem
than figuring out human heredity, and it took a hundred years to go
from Mendel's results with peas to cracking the genetic code.  On
the other hand the world's leading nuclear physicist, Ernest Rutherford,
said in 1937 that there was no hope of nuclear energy, and two years
later nuclear fission was properly identified as having actually been
observed in 1934.  You just can't tell.

	Anyway when Roszak quotes a ``dean of computer science at
Northeastern University'' as saying becoming computer literate ``is a
chance to spend your life working with devices smarter than you are, and
yet have control over them.  It's like carrying a six-gun on the old
frontier'', he has indeed found a fool.  All present computer programs
are extremely specialized and not very smart by human standards.  They
have no general knowledge of the common sense world, and they can't
even accept such knowledge from people.

	However, Roszak goes on to say that machine computation is
intrinsically different from human thought.  This amounts to saying that
humans can never understand human thought well enough to make a machine do
it.  That is a proposition that also requires proof.

	The present scientific situation in artificial intelligence
is roughly like this.  A number of important intellectual mechanisms
have been identified and understood well enough to embody them in
computer programs.  Others are identified but not understood.  What
we now understand is the basis of an expert system technology that
has practical importance.  How much practical importance is question
that a lot of people are betting a lot of money to find out.
While some computer scientists pursue
applications, others continue
basic research.  Recent discoveries assure us that progress continues.

\noindent Prejudices:

	Running through the book is the theme of establishment plot
as in ``What computer enthusiasts overlook is the fact that data
glut is no some unforseen, accidental fluctuation of supply, like
a bumper crop of wheat.  It is a strategy of social control, deliberately
and often expertly wielded.  It is one of the main ways in which
modern government and interest groups obfuscate issues to their own
advantage''.

	Also like many academic humanists, he confuses his taste with virtue.
Some of us just appreciate technology the way others appreciate opera.
We read about and buy new cameras, hi-fis and computer programs
beyond immediate needs the way others buy books.  Many also buy books
and go to operas.

\vskip .1in
\noindent McCarthy is the inventor of the LISP programming language and
the founder of the Stanford Artificial Intelligence Laboratory.
He has been Professor of Computer Science at Stanford
University since 1962 where he works mostly on giving computers
common sense knowledge and reasoning ability.

\vskip .1in
\leftline{John McCarthy}
\leftline{Computer Science Department}
\leftline{Stanford, CA 94305}
\leftline{(415) 723-4430}
\leftline{Social Security Number 558-30-4793}
\vfill\eject\end