COMMENT ⊗   VALID 00003 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.require "memo.pub[let,jmc]" source
C00034 00003	.skip 1
C00035 ENDMK
C⊗;
.require "memo.pub[let,jmc]" source;
.CB COMPUTER SCIENCE IN THE FAR EAST


	This is a report on a trip that included Hong Kong, Taiwan,
South Korea, Japan and Mainland China.  The visits were all brief,
and I am sure that I didn't hear of most of the computer science
activities in any of the places.  To most readers, the part about
Mainland China will be the most interesting, so I have made comparisons
wherever it seemed possible and appropriate
between that country and both Taiwan and the Soviet Union.

	In Hong Kong I visited for one day the Hong Kong Polytechnic
Institute and Hong Kong University.  I missed Hong Kong Chinese
University which also has computer science activity.  The greatest
interest is in practical topics such as operating systems and data bases.
There is much less interest in correctness of programs and analysis
of algorithms.  The professors I met were mostly British and Americans
under contract, but there were also Chinese including one student
working on a thesis on recursive programs at Imperial College, London.
The most interesting thing was a plan for a 300 terminal system for
student use involving several PDP-11/70 computers.  It seemed to
me that this was too many terminals for the computers, but a lot
depends on what languages and programs are allowed.
Hong Kong electronics manufacturers are beginning to look into the
possibility of independent manufacture of CRT terminals and similar
computer equipment rather than merely supplying products for
foreign companies.

	In Taiwan I visited the Academia Sinica in Taipei where an
Institute of Informatics is in preparation to join the thirteen
other scientific institutes.  As in the Soviet Union and Mainland
China, the Taiwan Academia Sinica operates scientific institutes.
They preferred a lecture Dialnet and LOTS to the more theoretical
lectures on mathematical theory of computation and epistemological
problems of AI.  They showed me a PDP-11 system with 10 or 15
terminals.  Although it was running at the time, no-one was using
it, presumably because it was vacation time.

	Next I visited the Tatung Institute of Technology which is
connected with the Tatung Company.  The Institute has grown up from
a trade school but probably hasn't yet reached a high level.  They
also preferred the practical lecture.  At the Tatung Company, I
saw a room full of girls typing the names and addresses of the
customers of the power company into the computer in Chinese characters
and a laser printer that was used to print bills also using Chinese
characters.  Input and output of Chinese characters is a current research
and engineering topic throughout the orient.
The equipment was from Fujitsu, but the brochures had also the name
of the Tatung Company, but I didn't determine what part they played
in the development or the adaptation of the terminals for Chinese
use.  The terminal included a keyboard about 30 by 40 in size with
eight characters on each key.  There was a numeric keyboard on the side,
and the typist designated a character by pressing the key on which it
was written together with a numeric key to choose one of eight.  They
were typing close to one character per second.

	The final place visited in Taiwan was the Tsing Hua University.
They also preferred  the practical lecture, even though my host there,
Professor Lee, worked in the U.S. on theorem proving.  He said that
interest in theoretical topics was just beginning.

	In South Korea I first visited the Korean Advanced Institute
of Science.  It is a graduate school under the Ministry of Science
rather than the Ministry of Education.  Former Provost Terman of
Stanford took part in an Advisory Committee that recommended its
structure.  They preferred the practical lecture also.  Artificial
Intelligence is in their catalog, but I was told they don't yet have
anyone to teach it.  Next I visited the Korean Institute of Science
and Technology which seems to be modelled on SRI.  It operates a
time-shared IBM 370 (or Japanese equivalent, I forget which) and has
remote users all over Korea.  Their director of programming did
what we call a FINGER command and was able to identify all the users
and the applications they were developing.
The Koreans were the most self-confident of any of the countries I
visited and were chafing on their dependence on foreign computers.
Perhaps they will be making some of their own soon, and they also
talked about making terminals.

	My final visit in Seoul was to Song Am University which is
run by American Jesuits.  Their computer expert, a Korean, had just
left them to go into industry, and a computer that their president
had begged from Univac was being installed.  No-one could say its
exact characteristics and their wasn't a handy manual, but since it
had 32K bytes of core, two fast printers and two fast card readers,
it seemed to be Univac's answer to the IBM 1401.  It was called the
9300.

	The computer activities in Japan are almost as varied as those
in the U.S., so I can give only a few remarks about places I visited.
Kyoto University has a dual processor Fujitsu M190, and this comes
to somewhat more than twice the computer power of SCIP (not counting
SLAC).  The most striking thing was the monstrous machine for handling
listings.  They are automatically chopped and go onto a continuously
rotating baggage carousel like those at airline terminals.  I think
the listing is made on command from the user.  In addition there was
a dynamic storage system for listings involving racks mounted on
vertical chains as in a paternoster elevator.  The whole system
clanked impressively and looked very impressive.

	The Research Institute of Mathematical Sciences at Kyoto is
getting a D.E.C. 2020 in February, and the Kyoto Sangyo University
has a 2040.  An attempt by Tskuba University to buy a D.E.C. System
20 was quashed by the Ministry of International Trade and Industry
on the grounds that it was inappropriate for a major government
university to buy a foreign computer.  This indicates that the
Japanese government is still protecting their infant industries.
The Nippon Telephone and Telegraph Company (NTT) has made several LISP
systems for PDP11's and also for Japanese computers at their
version of Bell Laboratories.  Goto's 
Tokyo University LISP 1.9 has many interesting features.  Finally,
I visited Tskuba University, where Shigeru Igarashi is a professor.
This a new university in a new city and is quite modern.  There was
a terminal room with 24 terminals for students, but I think they
were using a single language.  Another terminal room with 20 terminals
was available when I wanted to experiment with a LISP function that
had been mentioned by Takeuchi of NTT.

	In connection with a proposal to have a two week summer
school in AI in Japan before the AI conference including an opportunity
for the students to use LISP, Japanese facilities were discussed.  It
would be practical to hold the school at either Kyoto or Tskuba, but
both would have difficulty with housing and possibly with institutional
bureaucratic practices.  The practice of using student dormitories in
connection with vacation time conferences has not yet developed.

	The main consumer innovation in Japan since I was there
in 1975 is the coin operated laundry.  Now there are lots of them.
Prices are even fiercer than before.

	I spent a week in Peking as a guest of the Institute of
Computing Machines of the Academia Sinica.  They were interested in
both practical and the theoretical lectures but wanted the practical
lecture first.  However, the lecture on epistemological problems
of artificial intelligence attracted the most questions and discussion.
A professor from Peking University who formerly taught logic in the
former philosophy department was recruited as translator, and some
others from Peking University attended.  There is no AI work in
China apparently, but they want to start.  They plan to attend IJCAI 6
in Tokyo next year, and estimated that they might send 10 people to
the AI summer school if it happens.

	Since about 1967, the Institute of 
computing machines has built three computers, the most recent of
integrated circuits.  The most striking fact is that all the
components in the machine except the integrated ciruits (built
elsewhere in China) were built at the Institute itself - none
in factories.  The latest machine has 128K 48 bit words of core,
a smaller thin film memory, and does 2 million instructions per
second.  It has paper tapes, magnetic tapes (compatible with
nothing else), disks, a display terminal and printers for latin
and Chinese characters.  There is also a plotter built elsewhere
in China.  All this equipment is obolete in its specifications,
and they now regret this total isolation.  This was presumably
an aspect of the self-reliance policy.  As far as ideas go, the
technology seems to be all Western.  There is a preliminary
batch system for the computer, but mainly it is operated stand-alone
on a sign-up basis.

	I was shown an experiment in entering Chinese characters
with a light pen and CRT.  The CRT displayed possible upper left
corners of characters and possible lower right corners.  The user
selects the proper upper left and lower right, and the machine
displays the characters meeting these specifications.  The user then
chooses with the light pen.  The demonstration was not time-shared
with other uses of the computer.  It seems to me that it is fair to
compare this research project with the routine commercial entry
of Chinese characters observed in Taipei.  Taiwan is far ahead in
practice but depends heavily on foreign technology.

	At Peking University I was met by the President, and
he told about how the University is just recovering from the
sabotage of the Gang of Four.  They also built a computer
but not quite so independently.  Theirs is much slower than that
at the Institute.  We discussed sending Chinese students and Faculty
to the U.S., the research interests of their staff and visited the
library and the computer center.  At least two of their staff
have substantial interests in common with people in the Stanford
Computer Science Department and sufficient accomplishment so that
they would be worthwhile visiting scientists.

	At Tsing Hua University in Peking, I visited the Electronics
Engineering Department.  They designed a mini-computer
compatible with the Nova and built four.  Industry built another
100.  The computer uses a Bulgarian made disk unit that is
compatible with units taking the IBM 5440 disk.  When I visited
the installation, teachers were learning to use the machine in
preparation for teaching courses involving it.

	Both Peking and Tsing Hua Universities are just recovering
from the excesses of Mao's last ten years.  There were no students
from 1966 to about 1976, the entering classes to pass examinations
were in 1977, and the first graduate students are entering this
Fall.  Work on topics such as programming languages was proscribed
and not resumed until after Mao's death.

	At least in discussions with foreign scientists propaganda
is absent.  The only mention of Mao during my visit was of his
mausoleum as the last-mentioned of the sights of Tien An Men Square.
The only mention of the Gang of Four, except in response to questions,
was in the introductory speeches at the two universities; the
Institute of Computing Machines never had to stop work, and neither
Mao nor the Gang of Four was mentioned there.

	The only spontaneous mention of Taiwan was at a dinner hosted
by the General Secretary of the Academy.  In connection with the AI
Conference, he said that they had no objections to Taiwanese attending
the conference as individuals but not as a delegation from Taiwan.
When I said that all attendees came as individuals and that no
delegations from anywhere were recognized, he seemed satisfied.

	Pressing my luck, I asked what he thought would happen with
regard to Taiwan, and he said that China hoped to "liberate Taiwan
peacefully" and that a majority of the population of Taiwan wanted
to rejoin the homeland.  He had no further response to my expressing
doubt on this last point, but no-one seemed to get mad or freeze up,
and the discussion turned to other topics.

	On another occasion,
I asked an individual privately
how many people had lost their positions in connection
with the ouster of the Gang of Four and was told one at the
Institute of Computing Machines and four in the Academia Sinica
leadership.  These were referred to as the "little gang of four" and
as "rocket cadres".  It was explained that most people never agreed
with them so that their ouster involved only minimal personnel
changes.  Previous American academic visitors who explained
how the Chinese really and sincerely believed in Maoism should
re-evaluate their ability to divine the true opinions of people
in communist countries.  I have no clear picture of what people
are really thinking partly because I was never alone with one
person for long enough to bring myself to ask hard questions.
However, the much larger amount
of personal contact permitted now, especially by overseas Chinese,
should give a better basis than before for estimating public
opinion in various parts of Chinese society.

	A great desire to send students at all levels and researchers
to study in American Universities was expressed by everyone I met.
We discussed their sending particular people to Stanford, and they
said they would pay all expenses including tuition for the students.
The first students are to go to American universities this fall,
and no political conditions, such as normalization of relations,
were mentioned.
My guess, however, is that proposals to send particular individuals
will be subect to some political scrutiny and that there will be
more proposals than their limited foreign currency can finance.
However, it seems that they are putting a very high priority on
foreign science and technology contacts.  For example, when I was
taken to visit the Summer Palace, for my money the best tourist
attraction anywhere, we had lunch there, and it seemed to me that
every table had a middle-aged Westerner, sometimes accompanied by
his wife, and a group of Chinese hosts.
I hope we will do everything we can to encourage this - short of
selling out Taiwan.  They also want to buy American computers;
in particular the Institute of Computing Machines would like to
buy a D.E.C. 2020 in order to start work on AI.  I hope we will
encourage this too.

	Another sign of ideological relaxation was a movie.  I saw it
at the International Club, where there was a tape translating the dialog
into English, but I was told it was playing throughout China.  The
movie was made in Hong Kong, entitled "She Smiled at Him Three Times",
and concerned a famous poet of the Ming Dynasty who sold himself into
slavery in order to get close to the beautiful slave girl of noble
family who had smiled at him.  Even the slaves in the movie wore costumes
better than any clothing I saw in Peking.  Nothing farther from
Maoist ideology can be imagined.  (One American Chinese later told me
that the incident was real, but the poet sold himself in order to
get the girl as his fourth wife - in parallel not consecutively).

	China makes a better impression than the Soviet Union in
spite of its much greater poverty.  However, it is hard to say
whether this is partly due to the lessening of repression in the
Soviet Union that permits us to hear dissent directly.  When one
can visit Peking and talk to publically known dissenters, then
the comparisons can be made more precisely.  Let me grumble that
when I made this point before, I was often told by visitors that
there were no dissenters.  Well either the Chinese are lying now
when they say they disagreed with the Gang of Four or they were
lying then when they said they agreed.  The latter, of course,
seems more likely.  There may yet be a Chinese Solzhenitsyn, and
fancy historical explanations of how the Chinese don't value personal
freedom harm both the Chinese and ourselves.

	Returning to comparisons with the Soviet Union, the scientific
institutes did not seem to have the guard posts omnipresent in the Soviet
Union, and children came directly into the Institute, but I couldn't be
sure that I just didn't see the guards.  I forgot to ask, specifically.
Also, the stores seemed to have somewhat larger selections, and
there weren't the lines trying to buy up an item before it vanished.
I suppose this is a consequence of China's low wage policy that
keeps the supply of money below the supply of goods.

	As for comparing China with Taiwan, Taiwan has the advantage
both materially and in personal freedom.  Taipei was said to have
more than a million motorbikes, resembling Europe of the early sixties,
and it sure looked like it.  This is almost as many motorbikes as
Peking has bicycles.  (Calm down, eco-freaks.  Peking will have motorbikes
as soon as possible - at least they have publically regretted their
lack of freeways.  Taiwan has a freeway running the length of the
island).  Economically, Taiwan seems further ahead of Mainland China
than the capitalist parts of Korea and Germany are of the communist
parts.  One can ascribe this to the insanity of the last twenty years
of Mao's life.  Even in steel production, so much emphasized in
communist ideology, Taiwan produces five times as much per capita,
and China only plans to double its production by 1985.  Taiwan,
with a fiftieth the population, approximately equals China in
foreign trade.  Its exports to us are approximately ten times those
of China, and this year we will have an almost three billion dollar deficit
in trade with Taiwan.

	In discussions of computing in other Asian countries, I began to
mention Taiwan without exciting any special reaction, but we'll see if
I am invited to visit them again as was discussed.  There may be a good
chance that China has come or will come to regard Taiwan in much the same
way that they regard Hong Kong - in principle theirs, but for the time
being to be respected for its economic accomlishments and as a trade
partner.  We can encourage this by sticking to our position that
normalization requires a commitment that Taiwan not be attacked.
This will also permit Taiwan to avoid such dangerous moral builders
as demonstrating the ability to make nuclear weapons that
they might feel impelled to if U.S. abandonment caused an exodus of
capital.  One Japanese newspaper said that Taiwan is already buying
oil from China, so perhaps matters are moving in a peaceful direction.

	In Mainland China's favor, one might mention
the more comprehensive scientific interests of its
scientists.  This is presumably a consequence of its self-reliance,
which may mean that no scientific topic can be left to foreigners.
The self-reliance has had the cost that China can't win a technology
with the rest of the world and therefore many of its products are
obsolete by world standards.  Presumably, the capability developed in
the years of total self-reliance can be used effectively to implement
more moderate policies.  I did meet a dissenter in Taiwan who thought
that Taiwan would lose because of its social inequalities.  I don't
think he was much impressed by my stories about official limousines
and other privileges I had seen in Russia and my presumption that they
also exist in China.  (It turns out that they do).  Other Taiwanese are
bothered by the difficulties in getting permission for foreign travel.
Perhaps this shows that there is nothing in Chinese history per se that
makes them value personal freedom any less than anyone else.
.skip 1
.begin verbatim
John McCarthy
Artificial Intelligence Laboratory
Computer Science Department
Stanford University
Stanford, California 94305

ARPANET: MCCARTHY@SU-AI
.end

.turn on "{"
%7This version of
TRIP.REP[E78,JMC]
translated for printing (by PUB) at {time} on {date}.%1