COMMENT ⊗   VALID 00004 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	page
C00022 00003	.require "memo.pub[let,jmc]" source
C00029 00004	.bb Political issues
C00031 ENDMK
C⊗;
page
2	I would prefer a less governmental future than Licklider seems to forsee.

9 The assumption about programation "that, unless it is wisely planned and
well executed, it will be disastrous in the long run." is one that
requires proof.  As it stands it is merely an advertisement for the
profession of computer-society guru.

10	If full AI is achieved all the other considerations probably
wash out.

10	"...the government organizes all the capabilities of society..."
gives us conservatives heart failure.  Why should the government control
%2all the capabilities of society."%1

11	Scenario I ignores Dialnet - the possibility of getting rid of
all the special networks.
by having any computer capable of communicating with any other by the
ordinary switched telephone network.  There is no special reason to
believe that packet-switching is a must technology; it is more cost-effective
for some purposes - that's all.  ARPA puffery.

14	I see the development as different from multi-net and less
complicated.  Dialnet permits communication, and the AT&T offers a
variety of bandwidths in its switched services.  Very short
connect times incur very small costs, and both the switching systems
and the receivers react fast enough to make one second connect times
reasonable.

	The commercial and other transactions are not integrated with
the communication network, merely buy communication from it.  There is
almost no dedicated communication hardware, because commercial
relations are so widespread and so changable and so efficiently
served by the communications monopoly.

26	Importance of security is exaggerated.

26	Essentialness of speech recognition is exaggerated.

28	It is not obvious that everyone can't protect himself.

32	One way of protecting against snooping on funds transfers
is aggregation.  I communicate securely within my group, and
things I buy are paid for by the funds transfer from the group.
The bookkeeping within the group is private to the group.
I have access to a number of unidentified accounts and transaction
I.D. numbers for them with purchase limits.  When I go to a store
I use these numbers rather than personal I.D.  It seems to me
that external transfers to such accounts can be monitored so that
an extorted transfer can be traced later, and the owner of the
account charged with theft.

33	I don't believe fine tight control is necessary for stability.

34	The picture of government law as natural law is falsified,
because government law is used as a means of aggrandizement by
groups.  That's what politics is about according to many political
philosophers.  Moreover, the disease of politics whereby representatives
of groups identify the aggrandizement (not necessarily financial) with
the aggrandizement of the groups themselves is getting worse.

35	The traffic control system envisaged is one of the least
objectionable of Licklider's plans, but even here there is the alternative
of less government control and more facilities for communication
and co-operation between computers in the cars and keeping an
audit trail for the purpose of apportioning liability and blame
in the rare cases when something bad happens.

36	I agree that the national interest should be considered in
the export of computer technology.  In the main, we should confine
ourselves to making sure we get paid for what we contribute to
other nations' technology, and this may mean restraining some
who can make a profit exporting technology developed by others
and really belonging to them.  In my opinion, the export of isolated
computers of even very great capacity to the USSR or China doesn't
transfer appreciable military advantage, because there is little
advantage in bootlegging a few hours on a CDC7600 for a country
that has hundreds of BESM-6's even though the CDC-7600 is (say)
20 times as capable as a BESM-6.  There would be more problem
if the export resulted in the transfer of large amounts of software
and led to copying the CDC7600 in order to use the software.
Therefore, hardware exports should be monitored in terms of
the opportunities for copying the produce and quantitatively only
in terms of the total computer capacity transferred.
Fortunately, the use of computers for business purposes is
not easily adopted by communist countries, because their systems
depend for their functioning on widespread violation of the
official rules.  The rules can't easily be changed, because
they represent the equilibrium among the powers of different
groups and chaos would result from their literal enforcement
by a computer program.

38	I don't believe that "knowledge based education" depends
on the development of new knowledge base technology.  
Just putting the Library of Congress in computer file and
making it available - just by its present catalog and existing
bibliographies - to everyone who has a terminal in his home
will be a major revolution.
Further gains in cost-effectiveness of hardware are not required
although they are surely welcome.  A technology based on AT&T
present dial-up rates, the commercially available cheap imitations
of the IBM 3350, and the Lear-Siegler ADM3 "dumb terminal" would
be entirely cost-effective.



39	I wouldn't ask for much attention now.

40	Monitoring intake seems like nonsense, but each individual
might benefit from a private but continuous blood analyzer.

41 	Earlids

42	The network exists.  It is called AT&T and what is needed is
more and better of the same, and no more integrated with the data-
processing.  If this approach is taken, the standardization required
at this level can be left to AT&T, its regulators and its competitors.

The Common Business Language standardization is a different matter.

43 Goddamn it, the hobby use of computers is coming along nicely, and
the government has done nothing to foster it, and only a little
aimed at strangling it.  What is the evidence that the government
has anything to contribute?

44	The utility of packet switching is exaggerated, and it isn't
clear that present technology would support an all-digital speech
system.  At least new private small systems aren't all digital
and packet.

46	If the government alternative were to appoint Licklider as
Czar, it should stay out.

47	I see no reason why the present telephone wires connecting
each telephone privately with the central office and capable with
boosters of megahertz communication, can't be the basis of every
service except TV described in th?s paper.

48	Dialnet will obviate the need for any special action by
anyone except independent manufacturers to achieve computer-based
message services.  It is a bad idea to combine a clearly understood
service with vague ideas about data-bases, documentation services,
computing services and others.

49	It would be good to elaborate the "new level of responsiveness".

51	I agree substantially with what is said about the opportunities
for computer based politics.  But one can go much farther - with
the right to vote on any issue combined with the right to delegate
the vote for as long a period as desired, but take it back whenever
the citizen has his own opinion.

52	I think the importance of government in the metafunction roles
is exaggerated.  Certainly few of todays goals and agenda are the
results of previous government planning.  It seems to me that the
universities and the media gur≠s play a larger role.  The schools of
education that often impart the values that the next generation of
teachers impart to the next several generations of students are especially
important.  In fact one might easily argue that the McGovern candidacy
of 1972 is more closely related to what was being taught in teacher's
colleges in the 30's than to the ideas of any other group in society
at that time.

55	Well there has been about zero use of the ARPAnet for control
or spying, because the possible controllers are already overloaded
with information.  No additional raw data from the labs could possibly
be of use to them unless they had programs to understand it, and these
would require full AI.  Suppose the government appointed the Stanford
AI Lab to work full time at trying to control the M.I.T. AI Lab.  What
could be done by snooping in their files, beyond what could be done
by reading the reports they publish and with which ARPA is already
bombarded is trivial.  If we had the full AI needed to figure out
what they should really be doing from their computer files, we wouldn't
need them to work on AI.

55 I don't see the need for government regulation beyond that involved
in regulating natural monopolies, i.e. the communications network.

56	It is already possible to communicate by Dialnet between
computers all over the world.

60	It's really hard to see that the past development of computers
was centrally planned.  Except in the initial stages of the
development of scientific computers, the government hasn't been
the largest customer even though industries supplying the defense
department have been large customers.  None of the present generations
of computers or hardware were primarily government supported.  The
government role in computer development has been far less than it
was in aviation.  Before 1955 the situation was different.

63	The number of jobs is regulated by other servo-mechanisms
than introduction and disappearance of technologies.
65	No-one, especially Licklider or Shubik, has given evidence
that computer-based planning is likely to be of great importance
in the near future.  This is because the main problem of planning
at present and for the forseeable future is likely to be political
not technical.  Perfectly good hand-generated plans are not adopted,
because the many of the interests that defeat the plans depend
on not being explicitly recognized.  Thus lawyers never explicitly
dicker for plans that give rise to litigation, but a plan that
promises to abolish a major area of litigation will meet all
sorts of difficulties having no ostensible relation to this.
The same is true of plans that diminish the importance of Congressmen,
professors, governors, labor leaders, or data-processing managers
in corporations.

66	On the contrary, the free market avoids brittleness, because
the participants in the market find each others behavior
unpredictable even in non-emergency conditions.  Even more important,
ciiting the 1965 power failure as the prime example of brittleness
shows the unimportance of the concept - or at least of Licklider's
concept.  It didn't do enough harm to justify much planning to
fight that war over again except at the technical level where the
`planning has in fact been done.  Actually our society is the
least brittle in history because it has such great reserves of
material and manpower and communication that could be mobilized
to meet an emergency.

66	Well they can type.

Government initiative has done nothing so far and will do nothing
in the future.  Why do you suppose government activity is more
important here than it has been in past technology?  We don't
even need any new monopolies.
69	What kind of computerization and programation do we want?
I vote for fractionated vs. integrated.  Secure if this can be
done cost-effectively and without excessive standardization -
otherwise not.  Reliable as mother.  Definitely not designed
to serve a defined national or public interest.  I vote for
each citizen using it to maximize his own interests according
to his own conception of these interests.  The same holds true
of corporations.  Specifically, for the reasons enunciated above,
I don't trust Licklider to define the national or public interest,
and the probable definers will know far less than he.

70	It is more important that public spirited people
develop their own ideas and submit them to the judgment of
the market place.

Points: The government should put its documents and policies on line.
	The more the policies can be formally rationalized, the higher
the probability that an appeal against a policy can get to someone
authorized to change it.

Importance of security problem is exaggerated.

Common business language.


.require "memo.pub[let,jmc]" source;
.cb Comments on Licklider's paper

	There are three areas in which I am unconvinced by this paper.

	#. I don't believe that the government must do something
about computers in order to avert "disaster".  I simply can't imagine
what disaster Licklider has in mind; in fact his inaction scenario
is not a disaster - merely missed opportunity.  Moreover, I
believe that economics will make computer networks intercommunicate
even if the government does nothing to force standardization.  A forced
standardization now would be harmful, because many important uses
of computer-computer communication haven't been thought of yet
and may require different facilities than would become standard if
decisions were made now.

	I find myself in substantial disagreement with this paper,
although someone far from the computer field might find the agreement more
substantial than the disagreement; e.g. we agree that the computer field
can make great contributions to human life in the next quarter century.

	My disagreements are in two categories - technical and
political.  The technical disagreements concern what technologies
are most promising, and what developments are %2essential%1.
In particular, I think Licklider overestimates
the role of some technologies with which he has been associated.
The political disagreement is that I see a much smaller desirable
role for government than Licklider does.  This carries over into
the technical sphere also, because I am inclined to try to
devise ways individuals and groups can do for themselves what
Licklider wants the government to do for them.

.bb Technical issues.

.item←0;
.bb∂8"#. Networks, cable television vs. the dial-up telephone net."
Wired communication has always been a natural monopoly, and
this seems likely to continue with presently forseeable technology.
I prefer to minimize monopoly, above all government monopoly.
I also find a government regulated monopoly more desirable than
a government sponsored cartel, because with a cartel, the main means
of competition is legal and political efforts to expand a company's jurisdiction
relative to its competitors.

	Therefore, I prefer to devise ways of separating the natural
monopoly communication aspects of networking from the computation aspects
which are not natural monopolies.  Licklider seems to prefer networks that
integrate services and communications even at the price of combining the
well-understood personal message services with other data services whose
form and market are not presently known.

	I think Licklider underestimates the capabilities of the
present telephone network.  Consider that each telephone now has
a private pair of wires going to the telephone central office, and
that this pair of wires can transmit data at a one megahertz rate
if boosters are incorporated as in AT&T's proposed Picturephone
service.  Even present transmission speeds are entirely adequate for
almost all the applications envisaged by Licklider including reading and
writing.

	For this reason I have been pushing an idea I call ⊗Dialnet. 
Dialnet is a set of conventions that will permit any computers equipped
with telephone dialers to communicate with each other without prior
arrangement or joining any formal network.  The conventions provide at the
basic level for a program in one computer calling a program in another by
ordinary telephone and engaging in two-way communication with it.  At a
higher level, the conventions provide for messages to persons, file
transfer, and acting as a terminal.  Thus we see that just a little
ingenuity can obviate the need for the government to impose
standardization.  I think that Licklider overestimates the
difficulties in linking present networks; I don't see why they should be
greater than the already solved problems of linking telephone systems.

.bb ∂8"#. Certain technologies will be helpful, but we needn't wait for them."
These include speech recognition
.bb Political issues

#. The assumption about programation %2"that, unless it is wisely planned and
well executed, it will be disastrous in the long run."%1 is one that
requires proof.  As it stands it is merely an advertisement for the
profession of computer-society guru or a cliche of the times.  In
fact we can bumble along as at present for as long as necessary.  It
is a question of missing opportunities rather than courting disaster.

#. Here are some rather statist remarks: %2"...the government
organizes all the capabilities of society..."%1.  Why should the
government have access to all the capabilities of society, and
why should we expect it to organize them well.