COMMENT ⊗   VALID 00028 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00003 00002	.require "memo.pub[let,jmc]" source
C00004 00003	.CB COMMENTS ON MOSES'S PAPER
C00011 00004	.cb COMMENTS ON NEGROPONTE'S PAPER
C00012 00005	.cb COMMENTS ON DERTOUZOS'S PAPER
C00017 00006	.CB COMMENTS ON WINOGRAD'S PAPER
C00022 00007	.cb Comments on Licklider's paper.
C00038 00008	.cb COMMENTS ON VYSSOTSKY'S PAPER
C00042 00009	.turn off "{"
C00044 00010	.cb Comments on Bell's paper
C00053 00011	.cb Comments on Simon's paper
C00058 00012	.cb COMMENTS ON GILPIN'S PAPER
C00062 00013	.cb COMMENTS ON NOLL'S PAPER
C00067 00014	.cb Comments on Shubik's paper
C00071 00015	.CB COMMENTS ON ARROW'S PAPER
C00072 00016	.cb COMMENTS ON NOYCE'S PAPER
C00075 00017	COMMENTS ON EVANS'S PAPER
C00077 00018	.cb COMMENTS ON DENICOFF PAPER
C00083 00019	.cb Comments on Minsky's paper
C00094 00020	.CB COMMENTS ON PERLIS'S PAPER
C00096 00021
C00106 00022
C00108 00023
C00116 00024
C00120 00025
C00126 00026
C00137 00027
C00153 00028	.cb Counters to Weizenbaum comments on Future Study
C00157 ENDMK
C⊗;
.require "memo.pub[let,jmc]" source

.cb Comments on essays.

.cb "by John McCarthy, Stanford University"

	These comments were solicited (and paid for) by the editors of
%2The Computer Age: A Twenty-year View%1, edited by Michael Dertouzos
and Joel Moses.  On seeing them, the editors
decided not to include them in the volume.  Probably this was because
they disagreed with the comments and found them too sharp.  I didn't
learn that they were being omitted till I received a complimentary
copy of the book.

.CB COMMENTS ON MOSES'S PAPER
.item←0

	Let me confess that I am a rival recounter of the 
wonders of the computer in the home.  However, the wonders I would cite
are not the same as those cited by Moses.  Somehow I don't quite believe
in his.  Maybe it seems too much like the 1925 accounts of the future cultural
benefits of television.  Perhaps my predictions would be subject to
similar criticism.

	I have three general comments.  First, Moses sometimes
prefers to solve problems in ways that require a thousand-fold
reduction in manufacturing costs when they can be solved almost as
well at todays costs.  Second, some really revolutionary possibilities
are missed.  Third, the paper suffers from the common academic desire to
determine what is good for the public, give them that and withhold what is
deemed less good - rather than letting everyone choose for himself.

	#. At present a terminal adequate for all the non-pictorial
applications Moses suggests can be purchased for α$950 in kit form.
Since it has no large scale integration, equally good
terminals can be expected to cost half
that price in two or three years, because the terminal industry is
extremely competitive.  A private telephone line to a computer or
concentrator five miles away can be rented for about α$25 per month.
We are not in a position to estimate computer costs, because we don't
know how many subscribers a present minicomputer can handle.  Possibly
machines are now cost-effective enough to provide the services Moses
depicts at α$50 per month, but it is also possible that another three
to five years hardware development is required.  We contrast this to
the proposal to put what is now a α$500,000 computer in the home.
Pie-in-the-sky is fine but not in competition with pie-in-the-hand.

	#. Similarly Moses greatly underestimates the possibilities
of the ordinary telephone network to provide home computer services.
He estimates a page at 10,000 bits, 1000 to 2000 bits/second communication
rate, estimates 10 seconds to turn a page, demands page turning in one
second, and rules out the telephone.  However, 4800 bit/second modems
already exist for unconditioned lines and 9600 bit/second modems for
conditioned lines.  Moreover, a micro-processor can be programmed to
expand text that has been encoded to half its volume.  Finally, have
1800 bits/second into my home terminal now without any information
compression, and I find it a trifle annoying for reading some times,
but entirely acceptable.  Cable TV on the other hand is not widespread,
and general purpose communication is unlikely to become a major interest
of the cable TV operators in the near future, while libraries can be
accessible through the telephone system as soon as the information is
converted to computer form.

	#. There are really revolutionary possibilities for demonopolizing
publication.  At present a book, magazine or newspaper
is produced by an organization with
many hundreds of employees and capital in the hundreds of thousands or
millions of dollars.  When there are large public data files, it will
be possible for an individual to publish by himself.  If he is well
known, the public will have their personal library programs check for
his new writings.  If he is not well known, he will have to get his
writings reviewed by influential reviewers.  A correspondent in
Constantinople could get his dispatches on Byzantine politics before
his faithful readers without the help of any organization.

	If you disagree with me about the above, dear reader, just
think that with the new system you could express your doubts in
a file accessible to any subsequent reader.  As he reads a page of
this, he need only type "COMMENTS?" on his terminal, and there they
would be.  He can also type "REREBUTTAL?"

	#. Fortunately, the home computer users will have implemented
electronic mail using the telephone system long before the Office
of Technology Assessment or the Postal Service can decide whether
to forbid it in order to preserve the monopoly.
.cb COMMENTS ON NEGROPONTE'S PAPER

	The vision of future of the arts presented in this paper seems
only too plausible.  I certainly ascribes to the arts a smaller role
in forming the human spirit than they have had in the past.

	The computer can indeed bring new tools to the artist.
Unfortunately, the arts today are not in a good mental position to use new
tools or even their old tools.
Somehow the view that an artist is primarily a maker of metaphors
contributes to this situation.
.skip to column 1
.cb COMMENTS ON DERTOUZOS'S PAPER


	I agree with Dertouzos's fundamental contention that information
processing technology will enable us %2"to reverse some of the
impersonal and dehumanizing consequences of the industrial revolution"%1.
- at least to reverse what are perceived as such consequences.  We
shall indeed have individually designed shoes and clothes at a
price close to that of the mass produced product.
On the other hand I find the detailed picture rather implausible
for reasons I can't always quite state.

	My basic problem is that Dertouzos sees the various marvels
as straightforward improvements of present trends.  Therefore, he
tends to project present ideas to the achievement of these goals.
I think we are at least one scientific revolution away from the
level of artificial intelligence required to achieve many of the
objectives present in this paper.  Therefore, the future will be
more intellectually surprising than the paper suggests.

	Here are some specific disagreements, mostly minor:
.item←0;

	#. Like most writers, Dertouzos jumps to conclusions
about the relation between technology and employment.  Looking
at the past, American industrial productivity has grown by
more than an order of magnitude since 1900.  Unemployment
levels are similar, and their is no correlation between
specific advances in technology and the level of unemployment.
The economic servomechanisms governing employment have a much
shorter time-scale than technological advance.  They can handle
much faster rates of advance than are likely in this country.
For example, they can handle the ten percent per year advance
that Japan experienced during the 1960s.  The error is like
someone trying to predict the effect of a change in the gain
of a stage of a feedback amplifier on the output.  There is
an effect, but it is completely masked by the feedback loop.
Unfortunately, we don't yet know how to control the economic
feedback loop.

	#. Count me with those who believe there is nothing
qualitatively new in connecting computers together.  I think
one can prove mathematically that one computer can simulate
n-interacting computers at a speed of approximately 1/n
times real time.

	#. While "arithmetic atrophy" of some kind is a plausible
conjecture, I'll bet $50 that Dertouzos can't cite a study actually
observing it (written before December 1977).

	#. One false note in the shoe scenario is the card.
The home terminal and personal data files will almost certainly
precede the shoe-maker, so one's measurements will be stored
in one's home files and so will the results of new measurements.
.skip to column 1
.CB COMMENTS ON WINOGRAD'S PAPER

	The loan refusal scenario, the diagnosis of what is wrong with
the way computers are used, the statement of how the public regards
computers, and the proposed remedies are all unconvincing.

	As long as computers exist, clerks will sometimes blame
them, when they can't or won't do what a petitioner requests.
The loan refusal scenario seems vaguely untypical, and it isn't stated
whether the clerk is putting his customer on.

	Whether clerks have become less responsive and, if so, whether computers
have made matters worse, should be investigated not presumed.  I would
guess that record keeping systems that involve weekly or even monthly
tape-to-tape processing have made personal request harder to fulfill,
because the data is often inaccessible for inspection or change, and
I'll bet this
has been much more important than the use of numbers rather than
names.
The transition to on-line systems should be encouraged, and then we
can really put on the pressure to treat people better.  My favorite
courtesy is that no organization should ask anyone a question to which
it already has an answer, though it may ask for confirmation.
Direct terminal access by the public to business and government data
bases and policy descriptions and rationalizations will
make more difference than use by clerks of natural language dialogs.

	I don't really know what ideas about computers different people
have.  I suspect that most people don't think about computers much,
because so far they merely do more efficiently what has been done all
along.  It is just as presumptuous for us computer scientists to
invent the public as it is for uninformed people to act on imagined
characteristics of computers.

	I think Winograd magnifies the importance of the field
in which he (and I) work in making bureaucratic systems more
responsive.  The key problem is to handle routine matters sufficiently
automatically so that people with non-routine problems can get
the attention of someone with authority to help.  This has become
more difficult, because the loss of self-confidence by every kind
of authority means that one can always get a bit more by being
a pest - with or without the help of lawyers.

	As Winograd suggests, the goal must be a system in which
a person can interact with a computer system that understands
the policies and goals of the organization rather well and can
automatically arrive at solutions that today are bureaucratically
impossible.  The more often an automatic system can arrive at the
same solution the board of directors of the company would reach,
the higher the level at which special circumstances can be considered.

	The present efforts at natural language dialog seem premature
in that the words used go beyond the understanding of the programs.
For example the program that accepts %2"I must be in San Diego
before 10 AM"%1 almost certainly doesn't understand %2"must"%1 well
enough to know when to ask whether chartering a plane is warranted.
Research in understanding the common sense world is more needed than
more dialog programs.
.skip to column 1
.cb Comments on Licklider's paper.

	Licklider is right that the new technologies of computers,
communication and databases provide great opportunities to improve
government and its interaction with other institutions and with
individuals.  His proposed vehicle is a government organized
%2"Multinet"%1 combining communication, data bases, transaction monitoring,
and computation.  Unfortunately, the %2Multinet%1 idea
misses major opportunities to benefit citizens that can be implemented
much more quickly, and the proposals
overemphasize government, require co-ordination
of activities that can better be developed separately, create unnecessary
and harmful monopolies, and stifle individual and business initiative.  I
also disagree with some of its technological judgments.

	Licklider seems to envisage government as embodying the collective
wisdom of society and setting up institutions and laws that once %2"cast
into both legislation and software, ... may come to be regarded the same
way as physical laws are regarded now: irresistable, inviolable,
inescapable, part of nature, invariably enforced, not the kind of thing
one complains about"%1.  For good or ill, government plays quite a
different role in every society in the world.  Influencing law and its
interpretation is a major means of struggle for group advantage and
individual wealth, power and prestige.  Therefore, government law is
unlikely to resemble natural law in impartiality and constancy in the
forseeable future.

	A consumer is better served when manufacturers compete for his
dollar than when their lawyers compete for the favor of Congressmen,
regulators, or judges - even when all these officials are trying to act on
his behalf.  Unfortunately, some services are natural monopolies and
require regulation.  When a complex of new technologies becomes available,
it is important to distinguish what are the natural monopolies and what
can be provided by competing servers.  The %2Multinet%1 proposal lumps
them all together.

	Thus communication of raw information may remain a natural
monopoly, but devices connected to the communication network are not.  The
operation of particular public databases such as a national library, the
airline guide or a national telephone directory seem to be new natural
monopolies, although different databases can have different proprietors,
and methods of accessing the database can sometimes compete.  Computation
services and much so-called value-added communication are not natural
monopolies.

	Here are some detailed comments:

	#. The stated assumption about programation that %2"unless it is
wisely planned and well executed, it will be disastrous in the long run"%1
is unsupported by any example of what bad things will happen if
data-processing and communication continue unplanned.  To be a bit rude,
it sounds like self-serving nonsense on behalf of the "public policy
community" and people who want the authority to regulate.  Certainly
Licklider's %2Scenario 1%1 can't be called a disaster.  It just depicts some
opportunities missed if the government doesn't take them.  Since
co-ordination requires the authority to forbid independent initiative in
the area to be co-ordinated, claims that co-ordination is required
are not just truisms and must be very well justified.

	I suppose the statement, %2"recognizing the potential
significance of computer communication, the government organizes
all the capabilities of the society to develop and exploit networking
in socially as well as economically productive dimensions"%1 is not
intended to be taken literally.  No-one has shown that networking
has potentials justifying government powers that were hardly
asserted even during World War II.  Licklider should say precisely
what powers he wants the government to have, making explicit
the cost to society of giving the goverment such power.

	#. We have in Scenario 2, %2"The Multinet has supplanted the
postal system for letters, the dial telephone system for conversations
and teleconferences, stand-alone batch processing and time-sharing
systems for computation, and most filing cabinets, microfilm repositories,
document rooms, and libraries for information storage and retrieval"%1.
Some activities, like telephone communication, are natural
monopolies, and others, like providing computation services, are not.
Experience seems to show that when an activity is a natural monopoly,
it is better operated as a private monopoly regulated by the government.
Government monopoly is worse and so is a government regulated cartel
(like the airlines or interstate commerce).  The reason why the latter
two work out badly is again political struggle for advantage rewarding
the best lawyers and the best politicians.

	Several recent events raise fears of the cartelization of
computation and communication.  First, Atα&T turned over the %2Teletype%1
network to Western Union and agreed not to compete.  Second, IBM settled
the CDC anti-trust suit by
by giving CDC the Service Bureau Corporation and agreeing not
to compete in the area.  Third, the Government's anti-trust suit
against IBM may result in a non-competitive division of the computer
business.  Fourth, the specialized communication carriers are
trying to persuade the government to protect them from competition
by regulating the services ATα&T and others may offer.
Worse than all the above was a %2New York Times%1 editorial
proposing to give electronic mail to the Postal Service.
Electronic mail is not a natural monopoly at all,
because a mail terminal is an ordinary product.
A terminal that dials a recipient's similar terminal can be marketed today
at less than α$1000, the cost of a local telephone call is already less
than that of a stamp, and the cost of a coast-to-coast call after 11pm is
21 cents.  If the Postal Service achieves its goal of a 35 cent charge by
1985 for a first class letter, they will be wiped out by α$500 terminals.
Of course, through the wonders of government planning,
individuals may be forbidden to communicate in this way.
Moreover, electronic mail is most useful in connection with
other services performed by the same computer, and these services
depend on the particular circumstances of the individual or organization.

	The article suggests that home computers require government
"fostering".  Notice that as soon as cheap microprocessors made it
possible, a private home computer industry arose that offers dozens of
products through hundreds of stores and supports more than five
magazines - all before Government even contemplated fostering it.

	#. The article shares the common system programmer's
self-important obsession with secrecy and
privacy and exaggerates the possibilities for crime and other
misbehavior that inadequate secrecy creates.
It also puts the emphasis on preventing more and more tenuous
evils rather than on how to do positive good.  Thus
the Multics effort produced a non-cost-effective
system because of its obsession with security.

	#. Packet switching is a worthwhile technology, and ARPA
deserves much credit for developing it, but ARPA's natural tendency
to overestimate what it has sponsored has led to a neglect of
possibilities for using the already ubiquitous dial network.
Elsewhere I have proposed that much useful computer-computer
communication, including messages, file transfer and login
capabilities of ARPAnet can be achieved for a few thousand dollars
capital cost if ARPAnet like protocols for using a telephone
dialer are developed.  This proposal is called %2Dialnet%1.

	#. The greatest opportunities for improvement in government
that computers permit is to make government more accessible to
citizens.  Here are some proposals:

.subitem←0;
		&. Every government document that is required by
the %2Freedom of Information Act%1 to be publically accessible
should be kept in computer files accessible nationwide by telephone.
This includes the Federal Register, the Congressional Record, the
files on the status of bills (already kept by Congress
for its own use in computers),
court decisions and the status of court cases, environmental
impact statements, and other rationalizations of government decisions
and regulations.
%3Implementation of this should begin right away,%1
because organizations and individuals can acquire computer terminals
much faster than the Government can create the data bases.

	&. Licklider proposes to improve the Government's
ability to extract information from businesses and individuals
by putting monitoring functions in %2Multinet%1.  I think this has
great beneficial possibilities provided the information is not abused,
and burdensome requirements of keeping information in the
government's preferred form are not imposed.  But how about
not allowing the Government to require an individual or business
to furnish information the Government already has in its files?

		&. A much greater long range opportunity
is much harder to realize.  Every policy generating
and regulation generating organization should keep active a
question answering program about the content and rationale of
its policies.  The object is to answer automatically those questions
that can be so answered automatically and therefore make the policy
makers more accessible to people with questions that can't.
Each person responsible for policies should spend
a few percent of his time dealing with the questions not answered
automatically.  %2The ultimate goal would be that a person confronted
with a regulation whose application is unjust in his
particular case should always be able to reach a person with
the authority to make an exception.%1  The full achievement of
this goal will require major advances in artificial intelligence.
.skip to column 1
.cb COMMENTS ON VYSSOTSKY'S PAPER

	Dr. Vyssotsky's scenario of the development of business computing
between now and the year 2000 is a very plausible extrapolation of
present trends.  However, there are some possible discontinuities.
.item←0

	#. I see home computer terminals coming into wide use in the
eighties.  I also see the complete equipment
of businesses with terminals - to the point where there are as many
terminals as desks.  This can lead to systems that increase bureaucratic
efficiency by larger factors than Vyssotsky projects.

	#. Electronic communication has a second dimension which
will be especially important in communicating between firms and
in communicating with humans.  Consider communicating a price quotation
for three ring binders at α$1.77 apiece.  In the a predetermined context,
whether on paper or in a computer, the quotation may be represented
as a character string like "1ZB0461 00177", but this string has
no meaning beyond this context.
An equally formal representation of the same information might
be ((PRICE QUOTE) (VENDOR SEARS) (CATALOG NUMBER 1ZB0461) (DESCRIPTION
"THREE RING BINDER") (COLOR BLACK) (PRICE α$1.77)) which might be a
"sentence" in a %2standard business communication language%1.  I see the
widespread use of such languages as an important possibility
in the next decades.
The writers of the programs that generate and use information in such
complete sentences need know far less of each others businesses than is
required for the kinds of interfirm computer-to-computer in use today.

	#. Finally, I have the impression that the increasing
bureaucratization of
American business, educational, and governmental institutions is an
unstable phenomenon.  It may continue until 2000, but it also may reverse
sharply when many businesses discover that they can do with many fewer
people in bureaucratic jobs, and these compete successfully with others.
.skip to column 1
.turn off "{"

.cb COMMENTS ON FERNBACH'S PAPER

	With a few exceptions, Fernbach could have made the
same remarks ten years ago, and in several areas, he is already
behind the times.  The most striking of these is his speculation
that perhaps each scientist should have a terminal in his
office.  Our laboratory reached that point in 1971, and the
many other laboratories are in that position today.  In ten
years most students will have terminals in their dormitory
rooms.

	When Fernbach lists the sciences, he forgets computer
science.  Perhaps he regards it as a minor branch of physics.

	He is right that the Library of Congress should be at
the fingertips of teachers and students.  The biggest barrier,
however, is the capital cost of reading the information into
a computer file.  The file storage and the communication are
already cheap enough - though barely.
.skip to column 1
.cb Comments on Bell's paper

"The Social Framework of the Information Society: The
Coming Age of Communication and Control" - by Daniel Bell

	To an unreconstructed technologist, Professor Bell's article
raises more issues than I have time for or the editors will have
space for.  I shall consider two major ones.

	First I would like to state the case for the old-fashioned view
that science and invention are progress in themselves.  Namely, they give
individuals, groups, and society as a whole more tools for dealing with
whatever problems they imagine themselves to have.  Since all the old ways
are still potentially available, a well organized society should always
achieve a result no worse than could be achieved without them.
However, when
some people take advantage of a new opportunity,
any losers and even small winners will complain.  To whatever extent the
complaints are justified, the fault lies with the politicians, the
dictators, the WASPs or any other scapegoat you choose.  They might
equally well take unfair advantage of a non-technological opportunity like
the discovery of America.

	It is argued that technologies have momentum, and people won't
stop using them even when they are bad, but such complaints,
to the extent that they are valid, should be addressed to the
organization of society.

	Actually I would argue that new technology is not today the
main cause of change in our society.  Present social and political
changes are mainly
economic and social reverberations of the technological changes of
the nineteenth century.  The pace of technological change
affecting daily life has slowed,
because the mine of easy mechanical innovation cost-effective at moderate
incomes was worked out early in this century.  Think how few postwar
inventions we rely on in our daily lives compared to our reliance on those
that preceded 1920.
If no more inventions were made, our lives would remain exciting for at
least another hundred years by which time the world would reach a uniformly
high but dull standard of living.
(By that time the communist countries will have solved the problem of
orderly transfer of power by the time-tested method of becoming monarchies,
and the hereditary general secretaries will begin granting real but
limited powers to their parliaments).

	Second, it isn't made clear which social projections of the future are
considered to be founded on economic and sociological reasoning
and which depend on specific projections of likely technology.
Some of the conclusions seem to me to be founded on technological
projections which, while plausible, are not necessarily the way
technology will lead us:

	Thus a major projection is that the relative importance of
information handling industries relative to manufacturing will
continue to increase.  It seems to me that this could be affected
by two kinds of development.  First, there could be a new expensive
desirable invention.  Thus suppose a reliable quiet personal flying
machine that could land anywhere were invented, and it cost α$75,000.
Suppose further that it was automatic so that children and other
non-drivers could use it, and it was fast enough to permit commuting
from 250 miles.  It could very quickly come to dominate the economy,
because people would strive to afford it, and if the technological
facts were to dictate such a price, millions would be
employed producing them.

	Moreover, the growth of the "knowledge" industry could be
reversed by new discoveries in automating paper work.  Once firms and
individuals keep all records in computers and communicate electronically,
there will be an enormous payoff in making interfirm and interpersonal
transactions automatically.  The public will even demand that the
government bureaucracy stop asking citizens questions to which it already
has the answer.

	On the other hand, it could be argued that the growth of "knowledge
industries", especially bureaucracy, is not at all a technological
phenomenon but a sociological one.  Perhaps middle class
young people prefer information
processing jobs, squeeze into them, and work to expand their functions.
The demand for "relevant" work of the sixties can be regarded as a
demand for the expansion of bureaucracy.  At least the job of a
maker and enforcer of regulations seems to come closer to criteria
for "relevant work" than does any other job not requiring specialized
training.  It can be argued that desire to be a lawyer had a major
effect in expanding the area of lawsuits.

	Anyway, I would like to know Professor Bell's opinion on whether
the post-industrial society is a technological or a sociological
necessity.

	In the description of present society, the impression of crisis
seems somewhat overdrawn.  In particular, I don't believe in information
overload.  People are as good as ever at ignoring what doesn't affect
their immediate decisions.

	Finally, I think the word %2compunication%1 is ugly.  Moreover,
it presumes that communication and computation are inextricably mixed.
Because communication is a natural monopoly and computation is not,
prices charged for raw communication should be kept separate and regulated,
while the price and services of computation should be determined by
the market.
.skip to column 1
.cb Comments on Simon's paper

THE CONSEQUENCES OF COMPUTERS FOR CENTRALIZATION
AND DECENTRALIZATION by Herbert Simon%1

	I find myself in agreement with almost everything in this
lucid paper.  There are only three points of difference.
.item←0

	#. It is not obvious to me that modern society %2benefits%1
from the increased power of the central government.  If the fifty
states had more independence, there would be costs, and what they
are is well known.  On the other hand, our present centralization
means that fewer energy policies will be tried, and this reduces
the probability that some state will find a near optimal policy.
The recent nationalization of educational policy is really a
tragedy.

	#. A corporation or government agency may centralize or
decentralize according to its leaders' ideas of how to achieve
the organization's goals.  Its employees have agreed to
do what they are told within certain limits.
However, a citizen has certain rights to dispose of his person, his time
and his property as he wishes, and this often dictates decentralization
when centralization would better achieve even agreed goals.  It is not
entirely the prerogative of government or social scientists to
structure people's motivations so as to achieve a "national goal".
Subgroups of the population may also have rights of self-determination
so that a slogan of "states' rights" is not necessarily meaningless.

	#. Electronic voting from the home is too quickly dismissed
as beside the point.  Suppose that all public statements are
immediately available - suitably indexed so that, for example, on
reading the President's speech, a citizen can immediately ask his
terminal, %2"Who, in the opposite party, has commented on that?"%1.
Suppose further that every citizen has a right to introduce bills
in Congress, but bills need a some millions of sponsors to reach
a vote.  Suppose that votes can be delegated and redelegated, and
anyone with 500,000 votes can have an office and staff in Washington
and call himself a Congressman.  Suppose finally that a citizen can
cast his own vote or redelegate it
whenever he feels so inclined.  Such a system might not be better than
the present one, but it would certainly
be different, and it would encourage political participation.
I would like to see it tried, and I would bet that initial instability
would damp out, and then it would work very well.
Even in such a system, effective ability to define the issues, would
remain in the hands of small numbers of people.
.skip to column 1
.cb COMMENTS ON GILPIN'S PAPER

	Professor Gilpin looks for evidence of the importance
of the computer in world affairs.  Apart from the fact that
computers are one more cost-reducing technology he doesn't find
much, and I must confess skepticism about what he does find.
I believe that the international financial clearing system would
work without computers; it has many fewer transactions to handle
than domestic clearing houses.  I also believe the Arab oil embargo
would have functioned just as well without computers to keep track
of who was selling what oil to whom.
He then tells about multinational corporations,
and I have just one doubt.

	It seems to me that corporations can
function only when contracts are enforced, and multinational corporations
can function only when international contracts are enforced - either
by nations internally, by powerful nations or by an international
mechanism.  When a mini-state, whose GNP is a thousandth of General
Motors annual sales, shoots the Chevy dealer and confiscates all
the cars, there is nothing GM can do except stay away in the future.
On a larger scale, the OPEC succeeded in quadrupling prices, because
the West lost its moral conviction that the enterprise and inventiveness
that found the oil, found the markets, and developed the industry
entitled it to enforce the agreements on which the industry was based.
To the extent that contracts can no longer be made, the world must
revert to more primitive economic forms, e.g. barter.  The U.S.
may have to obtain its energy domestically even when investing
to develop foreign resources would be mutually beneficial.

	He then goes on to consider the military consequences of
computers and doesn't find very much.  I agree that up to now,
computers have been just one more technology.  High level artificial
intelligence will transform all power relationships, but we can't
now tell whether it is five years away or five hundred.  To make
an analogy with nuclear physics, we don't know whether it is now
1905 or 1935.
.skip to column 1
.cb COMMENTS ON NOLL'S PAPER

	I agree with Noll's argument that government regulation of
computing will tend to create cartels and prematurely
rigidify the structure of the information industry.

	Namely, it cannot now be known what combinations of home
computers, time-sharing, specialized communication, telephonic
communication, private networks, public data bases, private data
bases in public computers, and private data bases in private
computers will work out best for the users of intormation
services.  This is because it isn't
known what services individuals, businesses and other institutions
will want and what they will be willing to pay.  Any attempt at
regulation in the near future is likely to split off the service
to be regulated and require it to be separately priced.  Let me
give two examples:

	1. The %2New York Times%1 proposed that electronic mail
be a function of the Postal Service.  One may imagine that the
Postal Service would be glad to have it as an extension of its
first class mail monopoly and not otherwise.  However, a hobbyist's
home computer is capable of sending electronic mail to any other
home computer over the telephone network once
suitable protocols have been adopted.  Moreover, the business
use of electronic mail will be inextricably intertwined with
other inter-business communications.  For example, we envisage
Company A's inventory control computer initiating a dialog
with Company B's computer leading to an order for ten thousand
pencils if the price is mutually agreeable.  Such a dialog
may involve five interchanges of messages per second for a few
seconds.  I would bet that the service initially offered by the
Postal Service wouldn't to this economically, because they
wouldn't think of it at first, and it wouldn't be important at
first.  On the other hand, an attempt to write a law extending
the Postal Service's mail monopoly to electronic mail would
undoubtedly cover such communication unless someone thought
to exclude it.

	2. The present laws concerning credit data bases
will give troubles in interpretation as soon as people start
using computer files to store mail - as many users of ARPAnet
already do.  When is a remark in a letter that someone didn't
pay his bill to be regarded as adverse credit information
that must be found and produced when the individual asks for
his credit records.  Most likely, the people to whom the request
is directed won't even know of the remark in someone else's
correspondence file.
.skip to column 1
.cb Comments on Shubik's paper

	Professor Shubik is precisely right in saying
that the goal of computer modelling of social phenomena should be to
increase both democracy and individual freedom, and I agree with him
that there are great potentialities.
I wish he would elaborate his ideas on how to %2"improve the scope
of individuality"%1.
Unfortunately, few will be convinced by the
article itself that computer modelling will contribute much to these
goals in the immediate future, because
there isn't enough information about the actual successes and
failures of the computer simulations that have been attempted and
the prospects for the near future.
Modelling duels mathematically and solving the travelling salesman
problem by computer just can't convince a reader that society can now
be usefully modelled.

	This unconvinced reader will be further distressed
by the remarks in the introduction to the effect
that unless advances are made in modelling, society won't survive.
Thus
%2"This technology is a necessary ... condition for the survival
α... of modern high technology mass societies which wish to preserve
democratic and libertarian values"%1.  Is this just a routine puff for
the author's topic or can the author identify some
ongoing process that will destroy our society unless computer modelling
wins its race with chaos?  Weizenbaum (%2Computer Power and Human
Reason%1) also says that computers have saved or will save
society from collapse.  While computer modelling and other uses of
computers will improve our society, no case has been made that we can't
bumble along as long as necessary without much help from computer
modelling.  Except that we must adapt our technology to lower grade ores,
I don't see any sense in which our society is "riding a tiger".

	I also don't understand, %2"This is a raw data rich but
information poor society."%1 Compared to what?

	Why are intelligent machines closer than wise or
skeptical machines?
.skip to column 1
.CB COMMENTS ON ARROW'S PAPER


	I have no comment on this paper.
.skip to column 1
.cb COMMENTS ON NOYCE'S PAPER

	Dr. Noyce's projection of the development of digital electronics
is entirely believable.  Clearly the harware will support all the new
applications of computing envisioned in this book.
I would like, however, to ask for something
more from the hardware side.

	At present, and in Noyce's projection, large
scale integrated circuits are major design projects and require large
scale manufacture to achieve reasonable economy.
%2What is the prospect of developing a technology that can make
small numbers of complicated large scale integrated circuits economically%1?
Preferably, the equipment should be usable by small organizations.  The
The technology being replaced by large scale integration which involves
mounting integrated circuits on printed circuit boards is accessible to
small organizations and even to individuals.  It would be unfortunate
to lose this capability permanently.  Micro-processors help but do not
completely solve the problem.

	We understand the fundamental limitations in the performance
of semi-conductor circuits.  %2Are there corresponding fundamental limits
to small-scale manufacture%1?
If the manufacturing equipment can be developed,
the problem of economical design will also be solved.
.skip to column 1
COMMENTS ON EVANS'S PAPER


	Evans's discussion of the interaction of computers and
communications concentrates on a survey of the communication
techniques available and likely to be available for communication
between computers.  However, it seems to me that the
factors presently limiting computer-computer communication (more
properly process-process communication) are not mainly the
hardware provided by the communication companies.  More important
factors are 
(i) lack of protocols for message,
file, and terminal simulation communication over the dial-up
network,
(ii) lack of a common language for expressing
business communications,
and (iv) lack of a national file naming and program
naming system.  Alas, these problems are not yet being seriously
addressed.
.skip to column 1
.cb COMMENTS ON DENICOFF PAPER


	Marvin Denicoff does a real service in attempting a history of
computing, and his topics are well chosen.  However, it is one of the
first attempts and naturally encounters disagreement from other
participants and observers.  Here are mine:
.item←0

	#. The fifties were more optimistic years than
the seventies and not only in computer science.  However, the picture
drawn by Denicoff doesn't do justice to the varied opinions extant
at that time about the probabilities of success of various endeavors
in given times.  When asked for a prediction of how soon a research
or development goal will be realized, a scientist is bound to give
an optimistic answer if he takes into account only difficulties that
he recognizes.  Therefore, he should cautiously answer that if the
difficulties he sees are all there are, it will take ⊗n years.

	#. Denicoff's history of time-sharing assigns to little
credit to Corbato (who initiated CTSS) and me and too much to Strachey
(read his 1959 paper), Teager, and Fano.  Institutionally, it neglects the
IBM and NSF support of early time-sharing at M.I.T. and N.I.H. support of
time-sharing at BBN.  Administratively, Philip Morse and J.C.R. Licklider
also deserve credit.
Fredkin deserves credit for recognizing that time-sharing was feasible on
a computer as small as the PDP-1 and devising most of the hardware
modifications required to make it feasible.
The importance of MULTICS is exaggerated; it never produced cost-effective
time-sharing and neither did the IBM TSS effort.  Denicoff inexplicably
totally ignores the PDP-6 and PDP-10 computers which initiated cost-effective
time-sharing and whose descendants are still the most commonly used
time-sharing machines.  Their success compared to MULTICS and TSS is
due to starting with real problems rather than the pseudo-problems
of levels of access and segmentation.
Many hundreds of PDP-10s are in use compared to a few dozen users of
MULTICS and TSS.

	#. Artificial intelligence.  The Dartmouth conference was in
1956.  The history of artificial intelligence is even harder to write
than that of time-sharing.  Time-sharing is successful, and careful
interviewing and reading will trace the successful ideas to their
sources.  Artificial intelligence has not yet achieved enough of
its goals to make it uncontroversial which activities are contributions
and which are blind alleys.  Moreover, the theoretical foundation
of the subject is still so undeveloped that it can't even be
definitively said what the basic ideas are.  E.g. is natural language
understanding fundamental or a side issue?

	#. Software development.  I find this discussion -
mixing issues of management, language and data
structure design, and proofs of correctness - rather murky.

	#. Large file systems.  One suspects the author of
overemphasizing activities sponsored by ONR.  The pacing
element in large on-line data systems has been the cost
of disk file.  Now that disk files are cheap, there can
be movement away from monthly batch processing of keypunched
data to immediate processing of data entered at terminals
as soon as it is received with immediate feedback to the
person who enters the data.  This change is important for
social reasons, because much of the rigidity and mystery
of computer systems has been caused by the infeasibility
of interrogating and changing data relevant to an individual
in the middle of a large tape library.  The problem may
be important enough to almost justify outlawing tape-to-tape file
processing.
.skip to column 1
.cb Comments on Minsky's paper


	This paper expresses opinions about many subjects related to
artificial intelligence.  The opinions are based on Minsky's great
experience as leader of one of the most successful artificial
intelligence laboratories and merit respect for that reason.
Nevertheless, it seems to me that many are not sufficiently
supported by arguments presented in the paper itself, and some of
them seem to me insufficiently clearly formulated.  In this commentary
I will list some I disagree with, some I agree with (especially when I
have something to add), and some whose formulations I don't understand,
and I will try to give reasons for my opinions.  It is a measure of the
present weakness of the theory of intelligence that so many important
topics have to be treated in this way as matters of opinion.

	The two main sections of the paper are a history of the
progress of ideas in AI and a discussion of the representation
of knowledge.

	The events in the history seem correctly reported, but
I am somewhat doubtful about the picture of steady progress
presented in the historical section.  All the problems solved
are in miniature versions and haven't yet led to a convincing
theory of the phenomena involved.  Therefore, I think the same
problems will have to be solved again and again until such theories
are developed.

	Unlike the hostile critics of AI, e.g. Dreyfus, Lighthill
and Weizenbaum, I regard this as normal for so difficult a
scientific problem and no reason to be discouraged.

Creativity

	Can a machine do more than its designer?  No!  A machine or
program can do nothing in a second its designer couldn't do in a year if
he weren't so lazy - except in a few specialized arithmetic areas where it
might take the programmer a hundred years to do a second's calculation.  I
personally feel the shortage of time and mental energy very acutely and
think that if the present state of AI permitted me to really multiply my
mental efforts by even a hundred, let alone ten million, I would
accomplish great things.

	Nevertheless, I think that there really is a worthwhile concept of
creativity and that creativity is not just those things we haven't yet
programmed.  In my view, creativity is the introduction of a new element
into a mental situation - where new means that the element is not
constructed in certain standard ways from the what has already been
considered.

	My favorite example is proving it impossible to tile with dominoes
a checkerboard with the two white corner squares removed.  The idea is to
count the numbers of black and white squares that must be covered . I
consider this idea new and creative relative to the concepts that arise
directly from the formulation of the problem, even though the idea quickly
occurs to a mathematician reasonably experienced with combinatorial
problems.  Thus creativity is relative to a stock of ideas, and there is
such a thing as "easy creativity".  Progress in AI and in the psychology
of intelligence requires the scientific study of such easy creativity.
The beginning of such a study lies in formulating a mathematical logical
notion of what concepts are immediately available.


Mathematical logic

	Minsky underestimates the potential contribution of
mathematical logic.  The essential virtue of a declarative language
is that sentences can be made to have meaning substantially independent
of context.  Humans need them for transferring information between
people who know little about the workings of each others brains.  However,
%2context-free%1 meaning
is also important for communicating between processes in a computer
and for storing information for future use without having to recreate
the intellectual situation in which the information was developed.  Of
course, this independence is imperfect, and most writers have mistakenly
concentrated on the imperfections of context-free meaning - missing its
importance.

	In fact, I doubt that the present generation of "expert" programs
will have a long life, precisely because the information they contain
is insufficiently context-free.  I think it will often prove impossible
to use two or more of them in a larger system.


Science Fiction Scenarios

	Minsky takes science fiction ideas too seriously, because he
underestimates the extent to which the writers'
requirements to produce a story have distorted their imagination about
the possibilities of AI and robots.  They need conflict situations
between approximately equal antagonists in which important issues are
seen to be at stake and in which individual effort makes a decisive
and clear cut difference.  It is also necessary for the author to hold
out an idea from his characters until the climax.  Often this is something
that will occur to most people years before the situation arises.

	In considering the "HAL scenario", one must distinguish between
programs that give advice from those given operational powers.  One
should not give an ill-understood program operational powers in
an area where its actions could result in disaster.  Secondly, when
a program gives advice, one can distinguish between using it to
get ideas and following it trustingly.

	There are important problems in comparing the human motivational
structure with that of machines.  As far as I can see,a human doesn't
usually have an overall goal to which all subgoals remain subordinate.
A goal often arises as a kind of metaphor - it is analogous to something
that satisfied hunger or elicited personal approval, but then becomes
entirely independent of its causes and is pursued even in opposition
to them.  Secondly, the weight given to various accepted long range
goals depends on physical state, i.e. when a person is tired, his ideas
of what goals to pursue in the next five years are altered.  Finally,
the human motivational structure evolved under quite different conditions
from those of civilization or even a tribal society.  As a result, it
has all kinds of peculiarities.

	The result of all these considerations is that it would be
both difficult and inappropriate to make computer programs with
motivational structures like those of humans.  Obedience coupled with
the requirement that the human user be given a full picture of the
consequences of all contemplated courses of actions seems attainable
and appropriate.  As we understand more about intelligence, we shall
know more about appropriate control mechanisms for intelligent machines.
My point isn't so much to say that there will be no dangers or problems
but rather that they will be different from those imagined by the
science fiction writers.
.skip to column 1
.CB COMMENTS ON PERLIS'S PAPER

A good paper.  I can't at the moment think of any comments short of rewriting
the paper to conform to my own ideas.
.skip to column 1
.cb Counters to Weizenbaum comments on Future Study

``The question which appears to be asked only rarely is what pressing
problems this innundation of technological fixes is supposed to attack.''

Maybe no problems.  Perhaps they only give individuals some opportunities.
Since no sacrifices are being requested, it requires only the that
the gadgets that further computer development makes available, seem
worth their price to the potential purchasers.  It is hard to see what
"pressing problem" the pocket calculator fixes, but we have them.

Weizenbaum's desire that all be subordinated to fixing " the aimlessness of
everyday life experienced by millions" is somewhat dubious.  Perhaps
the millions would worry that Weizenbaum has some authoritarian prescription
should people under his influence obtain power.

"overly technological society".  How so?

National information policy (quote from Bell on p. 784).  I see no
obvious need for a National Information policy except of the most
limited sort.  The Government should pay for a computerized version
of the Library of Congress (or else franchise a private monopoly
to do the same).  The only reason for this is that the one-shot
conversion of world literature seems to be a natural monopoly.

"scientism" p.787, of which I am an adherent (at least as the
term is defined by some of its detractors) is not an extension
of positivism.

As to Vietnam, no amount of computers will fix a bad policy -
whether the defect in policy be cowardice (as I think) or
imperialism (as Weizenbaum thinks).  Admiral Moorer is quoted
as criticizing computers, but he probably had even more critical
thoughts about the President, Congress, the intellectuals and
politicians generallly.

Weizenbaum is using an example from a field (Vietnam), in which
most intellectuals agree with him, to make irrelevant points.
Lieutenant Calley was convicted.

p.795 - While no-one may fully understand IBM's largest operating
systems for the 370, it is known that they cannot do anything
but allocate resources on the basis of very narrow criteria.  They
cannot act like HAL.