COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Transcript of ACM lecture.
C00043 ENDMK
C⊗;
Transcript of ACM lecture.

There has been alot of talk about revolutions; in the automation
revolution the computer revolution etc. and their amazing effects on our
daily lives and so forth and I'd like to claim that mostly this talk is a
mistake in that so far computers have not been very revolutionary, but
they will be.  Now the first part of the proposition requires simply
comparing the effects of computers on daily lives with the effects earlier
in the century of the automobile and the telephone and electricity in the
home and home refrigeration.  I think it is in general true that in terms
of its effect on daily lives the innovations since World War II have had
less impact than the innovations say from 1890 to 1920.  They certainly
have had an impact on productivity which has made everyone richer than
they would otherwise be and they have had an impact on health but everyone
is inclined to take both prosperity and health as his due.  I think the
public is less impressed by the innovations since World War II and maybe
this is one of the things that supports the antitechnological movement
which I personally regard as, so to speak, one of the main menaces of
today.  However such as it may be, I think we really are going to get a
computer revolution affecting daily lives and I'd like to describe some of
the things which I see as some of the components of it.  

One of these and
I think the one most important that everyone sees is home terminals and
home computers.  Now I have always emphasized the home terminal aspect of
it rather than the home computer aspect of it and I think this has been to
some extent confirmed by the experience of computers in the home which is
that after all in one's daily life one has relatively little straight
computation to do.  However, I think that the communication aspects of
home terminals and their aspect of giving us access to the world's
information is vastly more important and will be important to just about
everyone.  There are many things that one could say about this and I want
to say very few of them.  First of all, let us consider the consequence of
having home access to all of the world's publications.  For example, we
may imagine the Library of Congress being available on disk file.  To give
you a measure of where we stand in that technologically I read an estimate
in Electronic News of the expected sales of the IBM 3350 disk and it's
immitations and this comes to about five times the amount of information
in the Library of Congress.  In other words, this is the number of disk
drives so the people who buy these disk drives could keep the Library of
Congress on file for the rest of us as a kind of favor.  Actually, the
main problem is getting this information into a computer accessible form
and it would have been very nice if Jordan Baruch had appeared here and
announced a project to do that.

	But the government doesn't seem to see it
that way yet.  I once tried to promote this with Ed David when he was
President Nixon's alleged scientific advisor and he told me that he didn't
have much power and I didn't realize how little he had; he hadn't seen the
President in the last six months or something like that.  He was just
about to lose his job. But anyway one of the thoughts was that at that
time the U.S. had a two billion dollar balance in India and people
couldn't figure out what to do with it and my thought was that they should
hire people there to type in the Library of Congress.  However they didn't
do that and they gave it away.  Now having immediate access to everything
would, I believe, be useful and people would like it and they would read
somewhat more and somewhat more freely than they do now; I don't think
that it would revolutionize our society or anything like that.  One thing
that might have a somewhat more drastic effect would be the fact that if
you have such a system then you get a much more effective freedom of
publication, namely to publish something is simply to type it into a file
and declare the file to be public presumably with some sort of price for
getting at it.  Now, of course doing that doesn't mean that anybody will
read it or even that anybody will hear about it because surely no one will
read all of the accessions to the file anymore than one would read a list
of the accessions to the Library of Congress as a primary means of
discovering what he is going to read next. So it would depend on various
forms of publicity, presumably being mentioned in things that people were
in the habit of reading, but nevertheless it would have a very powerful
effect.  Because at present in the price of a book something like 80
percent goes to pay the cost of publication, that is of physical
production and distribution and something like 20 percent goes to the
author and in a computer based system this could be reversed.  This would
have various effects.  A major one is that someone could make a living
with a much smaller public than is required today especially since this
small public would be able to program its access so that it would learn
about this thing quite immediately.  There would be somewhat less of a
mass media effect.  One of the other things that we might ask the
government is that there is this vast quantity of information that is
required to be available under the Freedom of Information Act but it isn't
really available.  It's available in reading rooms in Washington and
organizations with professionals to do it for them can extract them but to
the average citizen it's not available.  One thing we could ask almost
immediately is that they be required to make it available in computer
files that are accessible by telephone from anywhere in the country for
anyone who is willing to pay the long distance charges, not too serious if
you do it late at night from a terminal.

	I'd like to mention one other
thing in terms which I believe will have a substantial effect although not
on daily life, and this is the direct communication between the computer
systems belonging to different organizations, for example, different
businesses.  I thought about this in connection with trying to develop
something that I call a common business communication language.  I'm not
interested in the low level aspects of format of bits per second and
protocol to say that there is an error in the message or things of that
kind.  I'm interested in the semantics by which one business's computer
says to another, "What is your price and delivery conditions for No. 3
pencils?"  I should say I would set off on this by reading an article
about the world of the future.  This was a 1965 model world of the future
in which things were heavily computerized and in this world of the future
a little bell would ring or there would be a discrete beep and a clerk
would turn from her typewriter to her terminal and see on there, "We need
3,000 pencils.  Order them from Company B". And she would turn from her
terminal to her typewriter and type out a purchase order. This would be
sent to Company B where another clerk would turn from the purchase order
to her terminal and type in,"Send Company A the pencils". As far as I
could tell this was written with a perfectly straight face, being the
world of the future.  Now, let me tell you a little bit about the problem
of this common business communication language; it has an interesting
scientific aspect.  The syntax is totally obvious:  it should be LISP,
namely, each item whether it be a sub item or a major item should start
with the left parenthesis and end with the right parenthesis and the first
element of this list should be something that says what it is and then the
arrestor and the rest of the information which has to be given in order to
transmit this message.  It should have one other interesting aspect which
I got out of Norm Chomskey's book, "Reflections on Language". He said
human languages have it and other linguists dispute that, anyway I think
this language should have it, which is that the syntax never requires an
......  What I mean is this:  Suppose a price is called for; then $3.12 is
certainly an acceptable price. But also an expression designating a price
like, "in accordance with our contract", or "same as last week", or "the
price given to your most favored customer" or any such expression should
be allowed.  Now initially syntax of the language can say that the real
problem turns out to be the semantics. If you for example want to give the
possible semantics for delivery conditions then the simplest semantics is
none at all.  Now in other words you say that everybody knows how this
stuff is customarily delivered; we'll delete how it's delivered.  And this
might be a time and place of delivery but leaving the message blank.
There might be a message specified let us say by Federal Express or
something like that but then there are more complicated things that may be
contingent upon various factors unless they are on strike,in which case it
goes by the Post Office unless they are on strike; it may be contingent,
or it may be by installments or something like that.  And in fact when you
try to decide what you are going to provide in such a system then you come
to problems of natural language but these are not the problems that have
been mainly discussed of the syntax of natural language; they are problems
of the semantics of natural language. And I've been sort of led by this to
the point of view that maybe much of the work in putting natural language
front ends on computer programs is beside the point, or at least from the
scientific point of view it is beside the point.  Because this work
consists of making the computer say in English something that you already
know how to say in computereze whereas the most interesting problem in
making a computer use natural language is to say how do you make it
somehow say those kinds of things that are said in natural language but so
far are not said at all by computers?  So isn't that sort of an
interesting question because this problem of the semantics of natural
language arises even after we have already made our major syntactic
decision that everything is going to be lists of lists.  

	Alright, one
final topic on these revolutions; I should mention artificial
intelligence.  I'm only going to mention artificial intelligence because I
believe human level artificial intelligence is one or more conceptual
revolutions away.  In other words,in asking when we shall have artificial
intelligence is like asking in 1890 when shall we have access to the
source of energy that powers the sun?  Let us imagine that Lord Raleigh
who was studying the problem at the time had made his calculations
somewhat differently.  He calculated that none of the present sources of
energy were adequate to keep the sun going for as long as geologists would
have preferred that it be going.  It's contraction was only good for a
million years or so and combustion wasn't even good for that long and so
forth, so what he did actually is that he haggled over how old the earth
was, whereas what he might have said is well, there must be some source of
energy which is quite unknown to us which is what powers the sun.
Supposing that he had drawn that conclusion and that someone had said in
1890 well, let's get that source of energy.  Well I suppose that we're a
little better off than that in artificial intelligence but perhaps it may
well be that one hundred years from now people will say that we didn't
have E=MC squared yet in 1979.  And this makes a nice transition perhaps
to my other half of this talk which is that we don't know enough about it
but we'd really like to know what it's social consequences are and further
more we can't really get anyone to pay any attention and they quite
properly won't pay attention.  To tell you what I mean by paying attention
suppose you were somewhat undecided as to whether you would prefer Kennedy
or Carter or Reagan or Connally for president this next year and you were
to say "Well, what I would like is a statement from each of them on his
position of the social consequences of artificial intelligence". They
would laugh and then they would get some hack to prepare a piece of paper
and there would be your piece of paper and of course it would be quite
worthless and if artificial intelligence during the term of office of this
person were in fact to develop an importance they would certainly properly
pay no attention to this piece of paper that they had produced as a
campaign thing.

Now I come to the second half of this talk which is what are the, one
might say, moral or social responsibilities of the computer profession?
And you've probably been deluged with advice on this question and your
eyes are beginning to glaze over, I will perhaps try to attract you by
saying that my advice is different from other people's.  Now the sharp
point on this is when somebody says, well, we should take responsibility
for the consequences of the technical developments that we are involved in
and often you can come back with the counter argument, and I do, which is,
are you sure that you know better now what is good for them than they will
know at the time about what is good for themselves?  And I think the
answer to this is usually "no".  People who undertake to decide whether
other people should be provided with some technological opportu- nity or
not are taking alot on themselves and I think my general view is that they
shouldn't do it.  It is our responsibility as technologists to provide
society and individuals, and I guess I will put it more individuals than
some vague thing called society, the opportunities with some confidence
that people are the best judges of their own best interest.  Another
aspect of this that people are concerned with is the effect of utomation
on unemployment.  In Europe even more than the United States there are
articles about the "chip" and they don't mean potato chips and I must say
that most of them have been rather silly.  For example I read an article
on Saturday in the Irish Times by a man by the name of John Ely who spoke
of the chip and he said almost nothing specific about it but he said that
the government should plan for full unemployment.  So his idea was that no
one would have to work anymore.  Now it seems to me that we can defend
ourselves in this respect and that we should defend ourselves by the
remark that in the main these developments are like previous technological
developments.  I think that we have the technical knowledge and the
responsibility to inform the public that microcomputers are not something
that is going to revolutionize our whole society in the next two years,
that it will take a long time to introduce them into the many thousands of
areas of technology in which they will eventually be useful and that their
impact will in the main be like that of previous technological advances.
What one can say about that is that unemployment is a serious problem in
our society but it appears that as far as all the evidence is concerned it
has little to do in its larger aspects with the introduction of
technology.  One can look at it in several ways. One is over time.  If you
compared the situation with that of 1900 there is much more technology but 
no more unemployment.  You can take it across countries                  
and you say you were looking at it for a correlation between
technology and  unemployment, it might even be positive, namely, the
countries with the largest unemployment are those that have very low
technology, or some of those that have very low technology.  For example,
the Chinese have just admitted to having 20 million unemployed.  Then
another way of looking at it is to say that there has been an economic
cycle in which unemployment has fluctuated but no one has ever claimed
that that cycle is in any way directly correlated with any technological
cycle, that is of the introduction of new technology.  For example, the
crash of 1929 is not ascribed to the introduction of any particular new
technology.  Now, one could ask how could this be because we can see that
a particular advance may displace quite a number of people and I think
that the way we have to look at it is that the causes of unemployment are
somewhat of a mystery but they are an economic servo- mechanism of some
kind which we don't understand very well but that technology is only a
perturbation on this.  To give a technical analogy, it is like a feedback
amplifier.  Suppose you have an amplifier which has a factor of negative
feedback of one hundred then small changes in its amplification in the
amplifiers inside the thing have almost no measurable effect on the
amplification of the system as a whole.  Now another topic I want to
mention is that I feel that the lack of effect of technology on daily life
since World War II is unfortunate and that in various ways we blew it and
that we missed opportunities.  I will mention two opportunities, maybe
three.  One is that if we had been, to say something controversial for a
change, more vigorous about our nuclear energy program we would be in less
trouble today.  Secondly, having just recently flow on the Concorde, I
think its sure a great thing and that one arrives less tired than one
otherwise would and that we blew it in not pursuing our own project in
that direction.  Especially as the alleged reason has turned out to be
nonsense.  But I would like to come to a specific computer area where I
think we did miss an opportunity in that I believe that we should be much
farther along in interactive computing and timesharing than we are today
and that the opportunity was missed in the early 1960's.  Now, the people
who missed in the largest degree who bear the largest responsibility for
missing it are IBM because they simply did not take sufficiently seriously
in the design of the 360 in the early 1960's the idea that computation
should be almost all interactive.  I know they didn't take it seriously
because of discussions that I've had with IBM people at that time.  Now,
one needn't cause that to say as a matter of blame; it may well be that
with the information that was available at the time that they couldn't
have made any other decisions than what they made.  Nevertheless, they
could indeed have done professionally alot better.  One introduction of
technology or technological question that has some kind of moral component
I would say is the issue of interactive versus batch processing as one
that does have a certain amount of moral component.  Namely, part of the
unresponsiveness of computerized organizations is that fact that if you
have some piece of information that is on a tape and it is really only
accessable by making another run of that tape through the machine then
people will find various ways to be unresponsive to inquiries and to
request to change something and so forth and one might say that it should
be considered a good and socially conscious busi- ness practice to see to
it that all information that affects the organization's dealing with the
public is immediately accessable so that the organization can be
responsive to requests of one kind or another.  It would be nice to put
that positively rather than say that there ought to be a law requiring
people to do such and such.  Maybe the right way to do it would be to say
that organizatons could in their advertisements claim that they are in a
position to be responsive and some organizations that are are banks and
they do try to compete on the basis of responsiveness.  If you could at
any time get your bank balance say, in the middle of the night if that
happened to be the time of your curiousity by calling up the bank's
computer from your terminal and identifying yourself suitably, that would
be an additional benefit.  

One issue that has been very prominent has to
do with computer files and that was already prominent in the early 1960's.
When I was asked to write an introduction to the special issue of
Scientific American on information in September of 1966, I put in it what
I thought would be a "bill of rights" about files; the right to know
what's in the files and so forth and the right to get things changed and
so forth.  Whether there is any causal in that relation or not I don't
know, but anyway many of these things are now incorporated in law, and now
I'm a little bit regretful.  I think that maybe I shouldn't have said that
that if one is going to talk about, so to speak, absolute rights of one
kind or another, I would almost rather prefer that there be an absolute
right to maintain any files about anything that an individual or an
organization wishes to maintain rather than the maze of beaurocracy which
is currently developing in this area.  It's not so bad in the United
States as it is internationally.  It is my understanding that people        
fought about shipping information across borders and other such more or
less nonsense.  I think that it is the responsibility of people in
technological occupations, including of course the computer occupation, to
defend technology, not only their own but technology in general against
moralistic attacks.  Now, in case someone doesn't know the meaning of the
word "moralistic" it's a perjorative term, that is if it's commonly taken
as a perjorative term it means the false appeal to morality.  I think that
we have really lost alot by our failure to pursue various technological
ends such as the one I mentioned about timesharing.  Let me mention a few
others which are not in the computer field at all.  I think that we have
suffered quite a few casualties due to coal mining and sulpher dioxide in
the air because of our failure to defend nuclear energy against attack.  I
think that every time we waste money it costs lives.  If we, for example,
hold something up from operating/ as is the issue in California of the
Diablo Canyon reactor for a year at a cost of two hundred million dollars
then one can make a super crude calculation that that cost sixty lives.
What you have to do is to compare the incomes and the death rates in the
poorer and the more prosperous states and you get a calculation that if
you put money into the economy then people spend a certain amount of it on
being healthier.  Or conversely, if you throw money down a rat hole then
people do not spend it on being healthier.  Now finally I want to mention
a kind of side issue.  Maybe I'm abusing your patience by bringing up
these various thoughts that I've had, which I should properly put in
letters to the editor of the ACM, but I want to mention the famous IBM
antitrust case.  This has two aspects as far as the computer profession
and neither of them have been given any publicity.  One of them is what
are our rights in the matter and the way I see this issue is that the way
in which IBM and the computer industry in general does business is
presumably to be rearranged somewhat for the benefit of the public but we
don't necessarily have full freedom to do whatever we think we want to.
The owners of companies have rights too, and I don't know have a clear
intuition of where theirs end and the public's begin.

	But before we can demand our rights, we must decide on our
interests.  What are the interests of computer users and of computer scientists?
I think that the computer users have some interests in how the computer
industry is organized and let me mention two of them.  One is that I
believe that a reduction in secrecy is least in the interests of computer
science.  Whether its in the interest of all the computer users is another
question, but I have certainly felt quite displeased with certain
technological developments being kept secret.  It would be in our interest
to try to encourage a reduction in secrecy. The other is that what is
clear is that one of the very likely developments from the antitrust thing
is the creation of a cartel in the computer field and I think this would
be a very bad thing.  To give an example of what I mean is that some seven
or eight years ago or about that time, there was a settlement of an
antitrust suit between Control Data and IBM in which IBM agreed to sell
Control Data's business in a certain field and not compete in it.  I think
that such arrangements, that is, of division of the market, are in the
best interests of the computer companies, that is, that they will have a
much more stable interest if the market is divided up nicely between them.
Perhaps this is in the best interest of all the computer companies and it
is in the worst interests of the computer users.  I fear that this may be
the direction in which things may head.  I think our major interest is in
good hardware and software and we've had some rather, not only in the
computer field but in other fields, spectacular restrictions on
competition in the name of competition and I hope that it is one of the
respon- sibilities of the computer professionals to recognize such trends
in our own field and to act against them.  Well, as you can see I have a
different concept of the public interest than a number of people and I
thank you for listening to me.