Defining Perspective, INEL Computing Symposium (1991)

From Viewpoints Intelligent Archive
Jump to: navigation, search
[Applause] of course the the main thing about
technologists is that we're good at understanding easy things like science
and electronics and other kinds of engineering and we generally hate
complicated things reason we're able to
do our business is because the things we study are pretty easy complicated things
like ecology and human psychology are
things we usually try and factor out but in fact a lot of what the next 10 years
is going to be about and some of what this last decade has been about is that
human interaction human psychology the
fact that humans don't change anywhere near as fast as the technology does
there's no real evidence to say that we've changed much in the last 30 or 40
thousand years what has changed is our
our ways of representing the world we
got smarter by coming up with better representations but we have to deal with
those reputations using the same old brain that we had when we're out hunting
animals in the forest and as Nesbitt
once said the higher the technology gets the higher the Touch has to go high-tech
high-touch and a lot about what has been happening
in the last 10 or 15 years of dealing
with computers is that most of the successes that have actually been made have not been technological ones they
the technological ones were predicted very well in a paper by Gordon Moore in
the late 60s he was he later became a founder of Intel and his paper he was a
physicist decided to take a particular silicon technology he happened to pick a
really good one MOS and decided to see where it could go if you took it to its
limits so he sat down and calculated how
many electrons what is the minimal number of electrons you could actually
store in an MOS transistor it's about 5,000 I think his
answer was that he decided how big that transistor would be and determined that
it could be made and then he got interested in what kind of improvement
could we make about where we were in the late 60s to get to this minimal MOS
transistor they came up with a formulation which is now called Moore's
law and it was basically that the rate of progress could be a doubling in cost
performance every couple of years and
interestingly enough I don't know how many people have followed Moore's law
but as of last year Gordon Moore's prediction was only off
by 5% in fact he was too conservative so
there's an interesting example of thinking about the future not in an
extrapolated fashion because he wasn't reasoning by trends that were happening
in the 60s what he did was to take an image of a desirable future and then try
and understand what the rate of progress could be towards that desirable future I
think that characterizes how one turns visions into reality part of it is
getting a vision the most dangerous visions are extrapolate or ones because
they assume that the present is going to be carried very strongly into the future
and of course it is for the uninteresting future all the unde interesting things are going on right
now we're hot items 20 30 40 50 years
ago but the surprising ones are the ones that we really are interested in the
unpredictable or the unexplored future I
remember I was I was assist systems programmer at the National Center for
Atmospheric Research in the early 60s
and we were working on the heavy iron of
the day we were just just spent six months at Seymour Cray his lab in
Chippewa Falls helping to debug the 6600 we were going to get model numbers zero
of that machine and one day I asked the
meteorologist what the weather is going to be like today and he went and looked out the
window I said what's going on we're
spending millions of dollars of government and hence taxpayers money
we're simulating weather patterns all over the globe on the hottest computers
of the day and you go and look out the window and he said Allen you don't understand the the weather isn't like
other things if some farmer went out today and plowed a field this morning
that difference in reflectivity could be
easily enough in Colorado here to start some updrafts and those could lead to
forming a thundercloud and we could be having a storm in fact an oak tree is
often enough in one place and not another - cause that instability and I
was just looking out the window to see whether one of those was forming
of course our grand weather simulations were using 9000 points of data for the
entire northern hemisphere of the earth so the the grain was a little bit coarse
compared to ploughed fields and oak trees I took that I took that to heart
because it occurred to me that of course
the interesting future is also depends on physical principles there's nothing
mystical about it there's cause and effect operating but it operates in this
realm of extreme instability where little pushes here and there will change large things in the community and one of
the tricks in predicting the future is to homed in on those I'm actually
working still working scientists they haven't put me out to pasture yeah I run
a couple of multi-million dollar research projects at Apple but I got
into this business because of some
things that happened to me almost 20 years ago now and if we could start the
slides there my friendly agent so in
1973 we in 1972-73 we invented this
machine had a bitmap screen with 500,000
pixels on it had six MIPS of computing power and was both the forerunner of the
Macintosh and the first workstations most of the computing technology that we
use today that is called user-friendly came out at the development of this
machine and the user interface on it
looks something like this of course we projected it the right way
around but it included an included desktop
publishing it included icons and
included objects overlapping windows all
the things that we're kind of used to today and we had a pledge when we
started up the Palo Alto Research Center that we were tired of all of the bubble
gum technology that we had put together in the 60s which could barely be
demonstrated by the inventor I'm sure you have a few of those here and we said
at Parc fresh start new money new sponsor this time we're never going to
we're not going to do anything here that isn't engineered to work for a hundred
people so if we do a time-sharing system here it has to be able to run a hundred
users we do a programming language it has to be able to be learned by a
hundred people outside the group that developed it if we do a personal
computer it has we have to make at least a hundred of them so when we started off
on this we basically were adapting crazier ideas that we had in the in the
60s and trying to bring it to earth and we set up a production line and we
started building these machines and after a couple of years we had put
desktop publishing on it the same year we did this machine we invented the
ethernet and another guy Gary
Starkweather invented the laser printer to go along with it he didn't know how
fast to make it so he made it a page a second it was partly because the most
convenient Xerox machine he could find move paper to page a second and proved
impossible to slow it down so
he didn't know what resolution to make it so he made it 500 to the inch so the
very first laser printer that ever worked moved paper to page a second and
imaged it 500 dots of the inch just like this machine is still more powerful than
most desk computers because we didn't we sort of overshot because we weren't sure
exactly what to do so the the the point
of this story is that by 1976 we had
built some 700 of these machines we built so many that every person at Parc
had one and every secretary had one and we had been using them for some years
and we did desktop publishing and many of the kinds of things that you are used
to today and what happened is is in August 1976 Xerox decided not to take
this to product we were surprised because they didn't turn down a paper
design they didn't turn down a prototype they were turning down a system that at
all are even working for three years well we thought this is typical of
corporate stupidity so we went on
inventing other things we continued building these we actually built about
2,000 of these machines during the decade of the seventies two years later
I met the person who had made the big decision at Xerox to turn this machine
down as a product and I was shocked to discover that he was rather intelligent
even for a Xerox executive and that was
when I really felt bad because then I realized that this is not a question of
stupidity because stupidity of course could always be better informed this was
a paradigm problem this was a context
problem we were like Copernican x' at park who
had wandered into a ptolemaeus convention we shared similar words like
we used the word orbit but we meant completely different things about them
and so what was happening was the worst thing that can ever happen when you have
vision and it's happened many times this century is that the vision can sometimes
be so strong as to change the context and when it changes the context
it destroys communication has happened
many times in science you think scientists would be better at it but Thomas Kuhn wrote a book called the
structure of scientific revolution pointed out that most scientific
revolutions in the last 300 years have taken at least 25 years to happen and
the main reason was that you had to let the old physicists die off
he said it in a nicer way so the problem
here I think is number one how do you predict the unexpected future because we
know the interesting things about the future are going to be unexpected in that they aren't going to flow in a
simple way out of what we're doing now number two those interesting things are
likely to set up their own paradigms their own ways of doing things and one
of the things that is characterized for instance the paradigm shift from mainframes to desktop computing is that
almost nobody in the mainframe world as far as professional programmers and
managers and so forth made the shift most of the programming on the desktop
machines have been done by a new set of programmers and as I like to tell my
friends in the computer industry there's a yet a third shift coming along and
that is going to be done very likely by a whole new set of people and the reason
is that we just are set up I think even biologically set up to want to do the
rest of our life what we've learned as a teenager and in our early 20s when we
get competent we have a goal of competence that we should really get
confident when we got confident we tend to sit on the competence and try and
blend all new things to fit with the old confidence that doesn't work I don't
think it ever worked but it only if it did it only work in a very uninteresting
world in our world the biggest problem
with dealing with computers as I hope to show you in the next few minutes is not
being able to sit down and switch them on and stick a disc into the disc drive and fiddle around with the operating
system those things that are often called computer literacy that's not the
problem at all that's not computer literacy any more than literacy in the
printing world meant being familiar with the Bible because that was what was
first printed no literacy has is a much larger idea and one of the things it
means is fluency of course but even more
important I think is a willingness to try and find
out what the new medium of expression can actually do and instead of
automating old processes into the new medium which is something we always do
and we start off to try and find out what new processes the new medium
actually makes possible now this has happened many many times in especially
in the 20th century Edison invented the movie camera in 1895 and his his thought
was that people would not stand actually talked about anticipating the future his
thought was that people would not stand for things much longer than a couple of
minutes which happened to be the size of the reels that they had back then his
idea of an evenings worth of entertainment was to paste two or three of these reels together so you would get
six-minute films well I was a perfect anticipation of the age of television
but in fact what happened
is that for 20 years people thought of the movie camera as a stand-in for a
person in the audience like you looking up on a stage like here from a fixed
point of view and so they didn't move the camera it was a simulated theatre
it wasn't until 1915 and DW Griffith did Birth of a Nation that modern film
language was invented which involved he realized you could move the camera you
could zoom you could do tracking shots you could do montages and all of those
things were what constitutes the modern language of film we learn it without realizing it because we grow up in it
but when they first did a zoom a
tracking shot into a train approaching
people screamed in the audience because their brains were set up to anticipate
anything getting larger as meaning it was coming towards them and they had not
learned to relativize that according to
a screen we don't realize that because we relativize it at an age where we
don't remember that we've that we've done it so each medium has its own
language and its own literacy and a lot
of our survival in the next 10 to 20 years will depend on not having these
20-year lags in which we simply take old obsolete practice and try and implement
it in the new media of course the computer is one of the worst of these because of its ability to simulate every
other medium that we've had it's a wonderful thing about computers but it also has led to us sitting on the paper
document for many many years beyond its
merits and one of the things is going to be hardest for the world in the next 10
years is to get out of the paper document and I'll try and show you why
that's that's necessary so
let me try and make a point here how do
we think about the future one of the
ways of doing it that's not extrapolate Ori is to try and understand how we
extend ourselves into our environment we're the naked animal we're the ones
that have to have intermediaries between us in the world I'm using one right now
which is language we of course think of tools as one of the ways we extend
sometimes we don't think of mental tools like language and mathematics as as
being tools and to me all of these are extensions of the gesture language and
mathematics are ways of capturing difficult things you can talk about an
elephant without having to drag an elephant along with you and point at it
you can make a little noise that signifies elephant and you can take that
noise with you and our brains are set up to be able to make that substitution
sometimes too well but in any case the M
word for tools is manipulation capture
things use something like our hands move them around now another way we have
extending ourselves that we think of less often is we extend ourselves
through agents agents are entities that can take on our goal structure so a
simple agent is to get somebody to manipulate the tools for us usually it's
a human being sometimes we can train animals to do it and we can go to Hawaii
or go to sleep on the hammock while we've somehow convinced this person to
do our our goals for us I think that's really powerful about agents is that
they can deal with other agents so if you come up with a set of goals that are
very understandable and attractive and give it to some agents they can spend
these goals all around the world and
because of this the M word here is not manipulation but management we
manipulate tools we manage agents tools with thumbs something that we look at
and we manipulate agents are things that look at us and we manage
them or try to now of course when the mainframe came along it wasn't either a
tool or an agent it was something outside of human scale like a factory or
a railroad and of course we have
mechanisms in our brain for dealing with things that are outside of human scale namely religion so immediately when you
deal with something this is true with government it's outside of human scale
mainframe computers immediately start winding up with ritual people in special
costumes special places where things are done and all of the things that have to
do with now of course we don't realize that because we don't realize most of
the actions that we do are motivated by deeper things in our head but the in
fact are so the interest people had who
realized this when the mainframe came along was immediately to try and find ways to reduce at the human scale and
the two strongest ways of doing it are one to try and turn it into something
like a tool and the second is to try and turn it into something like an agent the
tool is easier and so we were more successful at that in the last 20 or 30
years and the next decade is going to be the decade of the agent now one of the
ways of predicting the future and evaluating some new piece of technology
you see is to try and understand how well it fits into the context we have of
using tools and the context we have of using agents if it fits well you can
make a very good choice let me give you an example here Xerox the story I told
you about Xerox is ironic in a very strong sense because not too many years
before Xerox turned down a 10-year lead in personal computing they had it happen
to them they were a small company in the 50s in Rochester called Haloid
they got enamored of dry copying they brought up all of the patents for
xerography and every other patent they could find they invented dry copying
once but it wasn't too successful so they just went out and called it offset
printing then they went back and took another pass
and after many years of work they had formulated a machine which they called
the 914 the first xerox copier but
they'd done something that I wish more companies would do and that is they'd spent so much on research that they
didn't have any money left to build factories or to market it most companies
do the opposite what they do is they spend lots of money on marketing and
factories to build obsolete and sell obsolete things whereas the Xerox people
did a better job they came up with something really great and simply forgot
one little detail and they didn't know
what to do with it so they decided well let's take it down the road to IBM in
Poughkeepsie back then and so I took at
the IBM and they said die BM take this machine we loved it so much we just want
to see it get out we'll just charge you a small royalty build factories sell the
damn thing you know we'll be just glad to see it get out there they put six years of their life into this of course
IBM did what every large company does when they don't know what to do that as they went out and hired some consultants
the consultant is a person who knows a hundred ways to make love but doesn't
have a girlfriend
so of course all consulting firms are also named Arthur for some historical
reason the the one that IBM chose is called Arthur D little in Cambridge
Massachusetts and they went out and did the typical extrapolate or E market
survey that he learned in b-school and a
year dragged by eighty months later they came to IBM with six inches of paper
having spent a lot of money and gone out and talked to a lot of people and their
conclusion was that there was no market for a plain paper copying period their
two main reasons one is that there was no copy volume to speak of in the United
States so if you extrapolate zero
and the other was that the nine 14 was
about 10 times as expensive to run it cost about five cents to make a plain
paper copy on the thing and they said nobody's gonna spend 10 times as much as
they do for a mimeograph e right now so
IBM said okay well we'll turn it down so
they turned it down this is around 1958-59 it's all
explained by the way in a book by a Xerox executive called my years with
Xerox the billions nobody wanted
by jaundice our here interests as the classic story he tells it in great
detail this turndown action after the
long wait so enraged the Xerox people that they went out and got their board
convinced to let them use their own life insurance as security for the first loans for the first factory for the 1914
as I say that happened around nineteen fifty-eight fifty-nine less than 10
years after IBM had turned down this device Xerox was held to be in restraint
of trade by the Federal Trade Commission for being too successful remember that
1967 they were forced to share their
patents of this thing that only they were willing to take the risk on now
how could I be M have done this world's
largest computer corporation lots of smart people and just like the smart
people in Xerox 15 20 years later how could they have done it well the answer
was that they thought they were a computer company they could not imagine that they existed in any larger context
for example in the 1920s and 1930s not
one railroad in the United States of America invested one cent in any kind of
airline or airplane building their
entire conception of what it meant to have improvement was to have better
welded rails and faster locomotives they could not imagine for a million years
that they're in the transportation business they thought they were in the
railroading business IBM could not imagine they were in the
communications business their only image of themselves was what their current
products were and this is one of the reasons why the Japanese are likely to
roll over us in a big tide this decade because most American companies take
their image from what they're selling now and have no image to rely on when
they start looking further into the future Japanese are completely the
opposite Yamaha makes really great pianos and they make really great
motorcycles and the corporate management of Yamaha has absolutely no image
problems no conflict of identity because what do they focus on not the
noun not whether it's a piano or a motorcycle they focus on the word great
their whole focus is on their process whatever we do let's make it great and
have fun doing it that opens up the world enormous Li it allows multiple
points of view to be used and multiple points of view are the way you get out
from under this single-minded notion of what we're doing right now is the
defining thing for our world my friend Marvin Minsky who is one of the
inventors of artificial intelligence likes to say you don't understand
anything if you only understand it one way understanding something one way is
one of the worst things that can ever happen to a person because it becomes exalted to something that is almost
impossible to change heretics used to be burned at the stake and the reason they
were burned at the stake is because what they said were regarded as lies nowadays
we don't burn people who say what we regard as lies at the stake but we have
other ways of centering them but consider this
almost everything that we have today was a lie 20 30 40 50 years ago somebody comes
along in a less enlightened age and we're talking about this kind of
technology that kind of technology they would have been branded as complete liars when Galileo discovered the four
moons of Jupiter could see four the four brightest moons and Jupiter in his
telescope and he took the telescope to Rome to show the Cardinals of the church
they would not look in the telescope because because of their worldview there
could be nothing in there to see because
their worldview said that there could be only one Center of motion and any moons
revolving around a planet was another center of motion that could not possibly be so they didn't look and that is the
way we generally are about dealing with these new things which basically come to
us as lies lies lies damned lies and statistics as Mark Twain used to say and
in fact real question about these lies is whether there can be true in another
context if they can be true in another context that's when you have to watch out because that context may be ready to
burst upon us so what we need to do is
find some way of taming this guy and
people started a long time ago so one of
the one of the rules are particularly the future is elevate your context whatever it is that you think you're
doing right now almost certainly can't be definer of what you are visionaries
try and find a way of defining the goals of a group in such a way that allows
them to do other things a few years from now now another way of predicting the
future which is a little bit more mundane but works awfully well it's also
not extrapolated is to realize that because of the stodgy nough society and
because of the bubble gum and bailing
wire nature of most inventions most of the technologies that are going to
affect us in the next ten years have already been in somebody's lab for ten
years in other words it takes 10 15 20
years to get these technologies out of the labs and they're sitting there for
anybody to see when I give these talks I always ask the businessmen in the crowd
if they have anybody on their staff just
one person in a multi-billion dollar company whose job it is to simply go
around and visit research labs around the world there are only 40 or 50 good
ones one person can cover them every 18
months or so not a single company I've
talked to has it even one person to do this to just know what the developing
technologies are going to be and because of this every time one comes along
they're surprised they had no reason to
be surprised it takes a long long time now how about how about here does ine L
have anybody whose job it is to just go around and visit mmm
so that's enough that's one way of understanding early warning now let me
show you a couple of examples now one of the things I can predict for sure in the
future I hope is that we will be going
from analog to digital I speak with
reference to our video system here which is extremely analog when we were trying
it out it tends to drop frames and stuff so I hope you won't be bothered by that
but just consider when you ever you see a dropped frame that there's absolutely no reason for that frame to be dropped
or for anything that rope remotely resembles noise of any kind in the
transmission of information around I've always believed that it was a moral duty
of any company that made anything having to do with television to build in a
small hand grenade with a 15 year timer
the job of after 15 years is up especially in universities which never
get rid of these things is that this little thermal bomb will simply melt the
thing down then they'll be forced to go out and get it get a new one so let me
show you a couple of trips into the past the first one goes a long ways back all
the way back to 1966
yeah let me pause this and see if we get an image there actually let me ask the
crowd here because this is a group of fairly technical people how many people in their office have 19 inch
high-resolution monitors for their computer systems ok still a smaller
percentage so might find it interesting with this guy who's there with his hand
on a mouse in 1966 had a high-resolution
2 page monitor in front of him and a mouse and a hypertext system this guy
was Doug Engelbart who if anybody was the father of what we call personal
computing today I think he was let me give you an example of what his system
looked like if it doesn't jump around too much
so I can say I'd like to go to produce but I liked go travelers they get big I
like to say one branch only and let me
look just that low and I see it oh I can say I'd like to see one line only I can
see what does the very thing that I can do doesn't work nice that I have here
so here I'm afraid I'll need a different pictures of you so here's what I'll do
with a picture drawing capability or so slight lamp if I start from work and here's of you guys seem to have to go to
pick up all the materials and that's my plan for getting home tonight but if I
want to I can say the library what am I supposed to check up there I can just
point to that you know I see overdue books ago there was a staying at that
name supposed to pick up the drugstore
mmm-hmm I see those thing alright can do
things if I will just say I'd like to interchange the Protestant and materials
bingo interchanging them very quickly
cancer and get interchange with partners they
do it all gets renumber yeah so that was the the first hypertext
system actually hypermedia as you can see because he used diagrams as well
work really well it's very fast that little segment was from a demonstration
he gave to three thousand people in San Francisco in 1968 and I showed this
about a year and a half ago to the four Western Regional Bell companies they
were having a conference because they just discovered they had a network
they work themselves up to a fevered pitch most of the papers in this
conference were how they could send data over this network and so I got up and
showed Engle Bart in is more to the
video that shows conferencing with video cutouts and a whole bunch of things that
are just still dreams today and they say yes that's it that's what we want to do
I said well it's very funny because Engle Bart in the early 70s took all of
this to ATT AT&T and tried to give it to them and they threw him out on his ear
because they didn't have any understanding what it was that he was he
was doing back then unfortunately now he is still alive and is getting some of
the recognition he deserves he wrote about this system in his first grant
he'd been thinking about it actually in the 50s and his first grant for this
system the first official document was in 1962
and it scribes most of the details of what we would like to have on our
computers today I managed to get some money from ARPA and get some people to
build it and a lot of the technology we have today is the way it is because of
this system not because he was able to convince anybody in business that they
should do it but because he convinced us who are the next generation graduate
students and we said yes we want to have something like that actually my first
slice please my first
my first attempt at this was to say well
gee Engelbart system really doesn't work very well it's on a time-sharing system
that's really a pain to program a time-sharing system together every user
the cycles when they actually need them and so I thought well well we should
have something that uses the silicon technology we just got from the space
program so another guy and I built this machine not the first personal computer
very early one a desktop computer for very special desks namely ones made out
of steel this thing weighed about 350
pounds and had a fan that was likened to a 747 taking on problem was you can only
get like maybe 16 or 32 gates and one of
these little dual inline packages back then and this whereas your desktop
computer may have 50 little black boxes on the PC board this one had some 725 in
multi-layer boards and it was a mess but in fact it looks surprisingly like
the machines that we we have today and we built one of these and tried it out
and discovered to our horror that only graduate students liked it
that's because graduate students like most engineers delight in the the
puberty right of passage of proving that they can memorize 500 stupid commands of
a poorly design system [Laughter]
admit it that's why graduate students and
engineers like UNIX so much well that
was a shock to discover that and while I
was trying to figure out why what was
wrong with this machine because it had one of the earliest object-oriented languages on it had multiple windows a
head pointing it had a few things that even looked a little bit like the
Macintosh on it the package together did
not gel in any way with the people we were trying it on who are non computer
users at the time while I was conjuring
this in 1968 I happen to see this system
at RAND Corporation one without a keyboard at all first we erased a flow
arrow out of the way so that we may draw a box in its place now it's easy T wants
to draw a box and makes one for him now it's recognizing his handwriting the
printing in the box is being used as commentary only in this case the box is
slightly too large so when we change its size this is where modern-day window
control came from literally from the connector to the box attached a decision
element to the box and draw a flow from it to scan we then erase the flowers
attached to the process post new area and move the box to a new position
this allows us to draw a new box
well that system made a tremendous impression I mean it only would stay up
for 15 or 20 minutes at a time he's implemented on a computer system that
was the size of this room slightly less
powerful than an IBM XT but this is for
a single user the I'm just just an aside
these tablets were invented at RAND Corporation the same year that the mouse
was invented 1964 and they manufacture them and sold them in small amounts to
people who are doing experiments just to give you a perspective for some of the
younger people in the audience we bought one of those tablets to play with it at
Utah around 1967 or so the tablet cost
$18,000 not the computer the tab the
computers cost millions of dollars to do these things on but just the i/o gear
this is not the display this is just the tablet it was 18,000 of pre-vietnam war
type money so you can think today to buy
think of using an input device today experimental input device today that
cost something like sixty five thousand dollars without attaching a computer to
it one of the reasons we haven't seen a lot of advances in the last few years is
that people are not building input devices that cost sixty five thousand dollars they're trying to make do with
the ones that cost a few hundred dollars and progress is not being made well
I'd called what angle Bart did personal computing and I'd called my little
desktop machine a personal computer but this was completely different this felt
like you're sinking your fingers right through the glass of the display and
touching the structures directly so I made up the term intimate and I made a
cardboard model back in 1968 wondering
what an in an intimate computer could big could obviously not
a desktop computer and it couldn't even be a suitcase computer 0.5 variations
per block is not portability so I define
portability as something as portable if you can carry other things - there's
always other things you're lugging around we actually loaded this up with
lead pellets to see how heavy you could
make it before people would not want to use it and so forth this this device was
not science fiction it was based on Gordon Moore's paper because it was very
easy if you believe the physics of Gordon Moore's paper that I referred to you could calculate just when in time
you could build one of those machines and the answer was sometime in the late
70s and the or the early 80s if you really wanted to you really pushed it
and in fact I went to Xerox PARC to design and build this machine I called
this an intimate computer and it still
stands there as something that is about to happen in in the next decade now of
course the reason it didn't happen in the 1980s the way it should have is because of this extra 10 years of people
being completely misled by the IBM PC
but if you've been if you've been reading the papers you may have noticed
that even IBM has now seen the light that worries me tremendously because
when IBM decides to endorse something it means it's at least five years obsolete
so when they when they started building large object oriented programming groups
at IBM I was fearful okay so this is a
paradigm part of this talk another way of predicting the future is to try and
understand whether this thing actually fits into a set of rules it makes up
itself there's three ways to invent something the the dumbest way is by
brainstorming that's when you get a bunch of people in the room and you try and have ideas very popular in
businesses and other organizations I'm told Robert Heinlein said a committee is
the only form of life with ten bellies and no brain
basically you can't think strong thoughts in that kind of a meeting so
any meeting that's set up to try and come up with ideas and press for a conclusion distrust it tremendously
because you won't get what you want stronger way is to be goal-oriented so
they teach in college and Business School find a need and fill it and the
problem with finding a need and filling it is that it is extremely difficult to
get people out there that you're talking to to articulate what their actual needs
are what they articulate are their wants and their wants are not the same as
their needs nobody wanted the Xerox machine but they needed it nobody wanted
the personal computer but they but they needed it and so the Japanese do not do
market surveys at all what they do is build come up with the best ideas they
can they build five or ten variations of them put them out in quantities of 5,000
and 10,000 and see when they put it in directly into a person's hand whether
they actually understand what it is this
is the problem with goal oriented behavior goal oriented behavior assumes
that you know something that you're trying to go towards now of course the
strongest way of doing things is to pick directions and learn as you go that's
not the same thing as a goal find something because you don't know enough
to know what your goals are until you actually get fairly close to them this is something that's very difficult to
explain to people and it's also very difficult to explain this third way of
invention third way of inventing is to invent something that creates its own
need all the big hitters in the 20th
century are things that the very existence of them created their own need
because they change the context in which people live personal computer created
its own need the Xerox machine created its own need pocket calculator
hewlett-packard had a wonderful
marketing survey done that showed that nobody would use pocket calculators of
course the same kind of reasons for this arts machine nobody was using them
wonderful story by the way I knew the guy who did the pocket
calculator hewlett-packard and he told me this that the whole thing happened
because Bill Hewlett wanted one and boy does that help and Bill Hewlett in a
meeting one day took off his shirt and gave it to his engineers and said make
it fit in this the pocket of this shirt
so right so Tom Osborne and the others
went off and they worked and worked and worked in they got this awfully good
thing called the HP 35 wouldn't fit in the pocket of Bill's shirt so of course
they went to a tailor
this tailor did such a great job that bill Hewlett never noticed that the
pocket had been enlarged
see what I mean about the human factor so when we look at when we think about
computing most people tend to think of it as a relatively continuous thing
desktop computers are smaller mainframe
computers and the tendency is is to try and do the same kinds of things on the
mainframes do but what I want to claim here is that that's not so at all in
fact we've had already two major
paradigms you can even think of three if you include batch in there so you can
think of the biggest shifts in computing have been the shift from batch to time
sharing the shift from time sharing to personal computing and the next one
that's coming hard on the heels of this one we're only halfway through the personal computer revolution when I say
halfway through she mentioned that two and a half years ago personal computer
gross worldwide revenues past all of the
multi-user mainframe and many never to
rejoin them so in fact just two and a
half years after the revenues were passed at about something like 38
million 38 billion per year the worldwide revenues in 91 are likely
to be somewhere in the 80 billions of dollars and in fact the difference
between if the if you take the next five
years of worldwide revenues the difference between the worldwide
revenues and desktop computing and in mainframe computing over a five year
period is approximately 650 billion dollars that is a lot of money and yet
even as this is happening a new revolution is taking place and this is
this intimate computer revolution fueled not so much by the integrated circuit
which is the forcing function of the personal computer revolution but fueled
by pervasive networks not just pervasive
networks but pervasive to the extent that many of them will be Wireless of
various kinds the
people who do estimates of this kind of thing wait do estimates of this is you
wonder who are going to be the users you don't worry about asking people whether they want one or not because they don't
know but the real question is to ask can this thing be something that will
amplify this kind of person and the estimates about what the personal
computer was going to do depended greatly on the kinds of user interface
that it had the estimates on the intimate computer have to do with
similar concerns and with the kinds of applications that can actually be put on
the things - whether the thing is going to be appealing or not the belief is is that the saturation of intimate
computers intimate computers are the kind of ones that you would have in a in
a pouch like this or stuck in your hip pocket maybe you have a pair of glasses
for the display continuously connected
in by if we're in this room it'd be connected using some sort of infrared
transponders and out in your car would be hooked into the cellular system
people are starting that I know of seven firms including one large Japanese one
there are heavy at work on these and will start seeing the first versions of
these in just a couple of years the estimate is that somewhere between 1997
in the year 2000 the worldwide revenues for these kind of computers will pass
personal computers will pass desktop computers that is not that far off
people are investing literally billions and billions of dollars into trying to
do something about the desktop many of it is fruitless stuff like trying to hook them up to the mainframe
now the mainframe if that's a temporary
measure maybe remember whenever you try and orbit around a large thing with
gravitational field you wind up doing what it is gravitational field wants you
to do the in order to actually make use of your desktop computers you have to go
to distributed servers in order to get the kinds of throughput and processing
that you need outside of your desktop if you stay with the mainframes for some
reasons of compatibility for some reasons of holding on to old databases
as though they're assets rather than liability here's one of here's one of
the incredible things about the government they don't allow people to depreciate their code all right we can
depreciate Hardware because everybody can see that it deteriorates but because code on magnetic disks is
relatively eternal the government does not think of this as depreciating but in
fact most of the code in the US has been obsolete for more than ten years I once
visited a steel company and looked at
one of their billing systems and I could see the punched card a counting machine
layout that they use back in the 40s in
this program it had never been actually redesigned it actually been
reimplemented on many different kinds of machines and you could see the way they
were passing information around this thing you could see where the 4:07
printer was in the shop where the 101
card calculating punches were located in a thing and every time we do that we
lose enormous Lee because we're spending maybe 80 90 percent of our energies just
keeping the the old dinosaur alive we don't get let the mammals get out there
and start competing in a reasonable way well let me just go through these three paradigms without belaboring the point
performance you measure on what is called path length I
that's the millions of instructions that are executed by the mainframe before you
get a response back at the terminal I wants to ask 4000 Macintosh developers
if they'd ever heard the expression path line could number had because you don't worry about
millions of instructions executed before you get back to the terminal you are the terminal so we have MIPS on
personal computers it's like those great old dodges from the 1950s you could lay
a 450-horsepower you could lay a stripper rubber a block long turns out
completely irrelevant because the MIPS is actually an exponential change will
have some 50 to 100 mitts in the standard IBM and Apple machines by 1995
and astoundingly will have an excess of a thousand MIPS by the year 2000 they
won't be architected quite the same way that they are now but the the myths will
be there we will be able to use every single one of them by the way there are
never enough MIPS and even giving us something like 500 times as many MIPS as
we have now on a desktop we'll easily use it up just in making the user interface better no because the issue in
the future in the next 10 years is not going to be MIPS or pathlength it's
going to be access we're going to be networked up now the question becomes on
this worldwide network that I can hook up to almost every node for people who
are publishing tools and data they might need what is the probability number one
I can find those things that will really help me because I know they're out there
- what is the probability that I can get
it sent to me 3 what is the probability that I can use this thing once it comes
to me those are the issues of the 90s they don't have anything to do with
hardware compatibility or anything else anybody worries about hardware protocols
about things they're dealing in the wrong layer because the kinds of things
that you want to worry about our communications protocols and making
those the same across machines types of
data I thought green and white print out
sheets were gone forever I can tell from
the Snickers that you probably have some but I actually discovered my dentist a
couple of months ago was still using these things and needless to say I
changed dentists immediately I was wondering what what other old
tools he might be using personal
computers of the Macintosh or Windows type extend into fonts and graphics and
in the next five years we'll be simulating all of the media we've ever
used on earth including those like hyper media which can't directly exist in the
physical universe interesting thing if you drop Gone with the Wind the book
gone with the wind to the floor the ink and the paper feel the gravity and the
impact but scart Scarlett O'Hara and Rhett Butler don't write they're
actually in a different universe a descriptive universe you drop a computer
to the floor the silicon in it feels the gravity and the impact but the
simulations in it don't the simulations do not have to have anything to do with
our physical universe and some of the more interesting ones like hypermedia don't the computer is that thing which
given in dimensions of data to deal with
can give us one more dimension and that one more dimension allows us to make all
interesting things appear at the same place in hyperspace almost not used at
all yep today integration painfully with
files I discovered to my horror year or so ago that only in the US only half the
databases in the country are relational databases and this is after 20 years a
relational base stuff so half the database in the country are still flat
files from the 1960s maybe 1950s and
relational databases were obsolete the day they were born because they in fact do not extend in
any strong way and people are struggling and struggling with relational databases they're very poor way of doing things as
we'll see later cut and paste in personal computers from any application
into any application and in the next few
years ActiveX components that you can take any component cut it out as an object and
move it to any other place and have it work in a reasonable way how do you
print big printers
personal laser printers it's interesting we it was one of the strongest things
that happen at Apple was that Steve Jobs
happened to be very important in getting certain products out the marketing
people decided not to sell the laser printer at Apple back in 1984 is just
when I first got to Apple now of course we knew they said it was too expensive
nobody would buy it same kind of thing and of course we knew
that people would buy it like mad because we had done desktop publishing ten years before but they didn't know
that so Steve instead of trying to reason with them simply went and
threatened each one of them personally and he had eleven point nine million
shares of Apple stock and was chairman of the board and boy it just worked
wonderfully well they all the people who were against this sort of melted away
into the woodwork and in less than a week there was a plan for selling the
laser printers now of course the bombshell here and people have been
predicting it for many years and it hasn't happened is that printing is going to go away a printing isn't going to go away
because of any ecological sound reason it isn't going to go away because people
want to get rid of printing it's going to go away because increasingly we're
gonna have two kinds of models that don't print out how many people here use
spreadsheets for doing forecasting instead of just accounting anybody here
a couple people do you print your spreadsheets out for that no why because
how many different spreadsheets do you have to print out all right the interest
is in all of the different cases that are generated when you change assumptions in them and that is just the
smallest part of being able to use simulation models for things instead of
using databases once you go to a simulation model the interest in the
simulation model are the kinds of things that can tell you dynamically and you don't want to print them anymore and
once you go to hypermedia where you have a hundred dimensions of
data connected with and one dimensions of interconnect that
allow all interesting things to be at the same places you're not going to want to print those out because printing them
out destroys them so right at the time when this crisis is approaching these
little machines that can handle these dynamic and multi-dimensional models
will come along and they will be the printers they will be the repositories
for all of this information how do you give commands remember and type on the
mainframe and of course in ms-dos which is a mainframe one of these things like
the people who didn't realize they could move the camera for 20 years because of
course if you have your own machine you can supply a lot more cycles to the user
to help them remember the Mac and Windows you can see in point and in the
future because this future system is going to be much more agent like taking
on our goals it's going to be ask and gesture kind of software one purpose
custom applications are the typical things on the mainframes that usually train a few thousand people to run them
generic tools on personal computers and the reason the tools are generic is
because you cannot anticipate what a million people need and you can't train
a million people so you have to come up with things like spreadsheets and
desktop publishing systems and other software that the end users can tailor
to their needs and that is going to change in an even bigger way when we
move from millions of people using computers to billions of people using computers people will actually be
placing together their own tools and a lot of that is going to constitute what
computer literacy is in the future computer literacy if you make any kind
of
ich ins built for you by others so you
have to be able to have some sort of understanding of perhaps a visual
language like the Macintosh that allows you to deal with the thousands of things that other people are doing but you also
have to be able to write and writing doesn't mean writing in a word processor
that's an imitation of the old meeting medium writing means being able to make
some of those tools yourself to be able to make things that are like the things
you buy an egghead software and get banded over networks that is a real
challenge so if I any L is going to be computer literate in the next five years
what that means is all of the end users who have machines on their desks she'll
be able to fashion tools of a certain kind for them in a morally required
professional programmers just as writing formally required professional scribe
that's the challenge
these things are paradigms are programmed in completely different ways
mainframes are programmed in data structure and procedure languages like
Fortran and COBOL they don't case tools
are very big over in this left hand side
of the screen and the reason is that case tools are an orthotic race for a
hopelessly crippled patient I have this image of Fortran on a
hospital bed with many tubes running out of it because people would rather keep
it alive than to to learn anything new
in fact much of personal computing is object-oriented which is a very
different way of programming and there's
yet another way of programming that's coming along for the 90s that is required when we go to pervasive
networks one of the ways to think of this is if you move the distance between
a and B is a thousand miles and you move
something like a piece of data a thousand miles to somebody else just to
ask yourself the question what else do I have to move with it in many cases you
have to move a programmer at least you have to move manuals of various kinds
because there are all of these things that happen what happens if the format's
change in the source of where you move the thing do you then have to go around
and find all the data in the old format yes you do there's a rule that says you
never find all of them nor do you find all the procedures that do the old way
of doing things so it's a very fragile it generates lightning bolts objects are
much nicer because when you move one of them a thousand miles it takes everything important along with it
automatically all of the procedures that deal with its intimate representations
so objects factor out the desires you have to use them from the way they use
to represent themselves internally but
they still require a little bit of understanding at point B at least at the protocol level and our belief is that in
the next 10 years we have to move to a kind of module that we can send around
networks that does not require any human hands to touch it when you receive it
the active Ries it's like moving a shoulder muscle cell from your shoulder
and putting it down in your thigh you shouldn't have to reprogram the thigh muscles down there to accept it
that cell has enough knowledge about what it wants from the environment so
it's able to reconfigure itself after the journey and it's that automatic
reconfiguration that that in its knowledge of what it's about and what
it's trying to do that characterizes this as an agent or anyway of sign might mention that agent
oriented programming by the way is now ten years old object-oriented
programming is now more than twenty five years old so it must it should be on its
way here any minute now
how big is a big program this is why you should be interested in this I don't
believe that people should change programming languages just because of fads or styles but a big program in the
Fortran COBOL arena is a million lines of code maybe more equivalent kind of
thing in an object-oriented program it's about a hundred thousand lines about a factor of ten less and quite remarkably
the similar program in the agent oriented style which is kind of a
rule-based expert system way of doing things is much much smaller than that
often a factor of eight or ten and this
change is like thinking about architectures we're using the same old
bricks on the computer but we're using them in different ways the Greeks were
smart but they built rather small temples made out of posts and lintels
because they didn't know of any other ways of organizing the bricks quite
remarkably if you want to build a Gothic cathedral that's mostly windows in
mostly empty space inside it doesn't
require any more bricks than an average Greek temple does it requires a different architecture for organizing
these things this is like object-oriented programming and in fact you can go even further you can turn a
corner as Buckminster Fuller did even
though it had been known for a hundred years or so that some hundred fifty two
hundred years ago people suddenly realized with a shock that structures
most structures that we build things out of our much stronger when you pull on
them rather than when you push one if anybody here actually knows why any
sense you know when you push on things the remember they're all this heat there
at room temperature so the molecules are vibrating down there what happens when
you push them closer together they get hot in fact when you push steel together
it will start flowing long before it if you start pulling on it which makes
it cold and you give the molecular bonds a chance to do something for you so
fuller was the one who really fully realized that if you could build
structures in which most of the structure was pulling rather than
pushing you could build something that for instance that would go although would cover all of the Gothic cathedrals
on earth and out of about the same
amount of material cubic material as the Gothic cathedrals or the postable
entails this is one way of saying point of view is worth ad IQ points right our
whole ethic in this country is to work harder when we're in trouble but in fact
all of human progress has come not from working harder but from working smarter
every time we change the context in which we choose to solve problems in
we're made much stronger because we have no way of making our IQs any any
stronger so so technology is easy to
predict all at aw shucks I think this
will work this is our right angled right
angle view if you ask somebody for a screwdriver and they gave you this you'd
be upset because you say no I I want the
user interface because it's only the user interface plus the function that
gives me the tool and when I was making
this slide it got me to think about the screwdriver in a new way I thought God
that is a terrible design for a screwdriver just think of it the
mechanical advantage is the ratio of the diameter of the handle to the diameter
of the shaft that's small leads down pressure if you try and manipulate it
like this so you get a lot of friction on it it will slide off if you
manipulate it like this to get down pressure you don't get much purchase on it this is a terrible design for a
screwdriver this is the ms-dos of screwdriver designs
now my hobby is building musical
instruments I have a book published before the French Revolution more than
200 years ago that shows the tools that they used for musical instruments back
then and in there is a beautiful engraving of a screwdriver that looks just like that so this is a tool that
has been used almost uncritically for more than 200 years probably 500 years
back to the time when they invented screws that it's not been thought about
now what should the screwdriver look like well it should have a as wide a diameter in the handle as possible
you're going to use it from the top you need as much friction as possible so
what should it look like it should look like a ball is anybody seen a screwdriver it looks
like that yeah in fact since I started giving this this part of the talk I've
received some thirty eight ball handles screwdrivers most of them most of them
of different design which is really great but apparently tracing it back it
was about eight or nine years ago that somebody went through the same process and said you know that is a terrible
design for a screwdriver should actually look like the person's hand so the moral
of this story is you need to not only need you need a user interface but it
has to fit into something that the human has so if you want to communicate
instead of just manipulate you have to fit to something that the primate has
we're primates and we have to have something to do with how the mind works
now the problem with how the mind works is that we don't actually understand it
in a very strong way we think of ourselves as a seamless fabric of
eyeness there's AI me and myself and
we're all the same person and I can remember back when I was a kid and all that stuff but in fact the whole
situation is much more complicated than that give you an example here take a
look at this take a look at this upside down face here and see if you can see if
there's anything wrong with it what what about the teeth
upside down okay what anything anything else about this picture smile hole smile
what else no fear turning your head I'm watching you okay so guys okay good
so let me tell you let's pretend we're
at one of your training centers or in almost any school in the United States
most of these places believe knowledge is a fluid that can be moved drop by
drop from the full teacher vessel to the empty student vessel most of that
knowledge is encoded into language and the idea is if I tell you something that's true you're supposed to
immediately learn it and understand it so let's test that out
I'll give you an English sentence about this picture what we've done here is
we've taken a picture of a young girl her eyes and math ID down put them back
into the picture then turned the entire picture upside down that sound right
okay so having said that it should be quite prepared for what it looks like
right side up
I'm gonna turn this back because I've discovered nobody will listen to me when
it's on the other way
it's great to come up here to Idaho where nobody has seen this before I've
been showing this for 10 years and I've never found a better one and some of my
audiences and other places are getting tired of it but I like to throw it in it
because a couple of things to realize from this one is that even though I gave
you no additional information of any kind when I showed you the picture your reaction was completely different and
the reason is it was processed in a completely different part of your brain there's no seamless fabric of I myself
and me going on here instead we actually have modular mentalities when I gave you
the English sentence that went in through your ears and went over to the left-hand side of the brain where
languages processed turned into a proposition and you agreed that that was
probably the case they had a high probability of being true so everybody
just sat there when I went to the image that image went in through your eyes to
the back part of your brain and under the right-hand side of your brain where you Pupp sorry listen you might wonder
why am I looking that at it when he puts it on the screen the answer is that it
is biologically encoded into most mammals brains so if you try this
experiment out on a dog I suggest you pick a small one [Laughter]
this is not this is not a learned response it's actually built it's
actually built in there so when the
image comes in it is processed in a different part of your brain and your
whole psyche your whole consciousness comes to a completely different
conclusion which is startle met and so forth this is why McDonald's hamburgers
does not run print ads saying if you eat a McDonald's hamburger you'll become a
better-looking person right because it's
absurd in English but instead what they do is through run ads that show really
good-looking people eating McDonald's hamburgers and that information is processed in a completely different part
of your brain and you come to a completely different conclusion it's a weaker conclusion than you do with
language but it's inclusion that works extremely well for advertisers in fact
we have a part of our brain that thinks
things are more true if they rhyme so
our order to say to you a Big Mac and fries will slender your thighs
[Laughter]
you'll remember that - anybody have an
idea of why you might think things are more true if they rhyme and remember
them more easily any speculations
well association but why why why should
we be good at remembering rhymes patterns why should we be good at me see
the thing well we've only had writing for about maybe 10,000 years at the most
and for most of human existence we've
had to remember what the previous culture knew in terms of stories and
those stories were in given in terms of rhythmic schemes and rhyme schemes and
in fact I think that the survivors of
all of this namely us are here partly because we're able to remember things
well in terms of stories and in terms of rhymes and we seem to have that built in
from the from our earliest birth but the
point of this thing is that what we have as a mind is kind of a patchwork quilt
children deal with the world in very
comprehensive ways but different ways as they get older they learn about things
by grabbing them and eating them early
on they go into a visually centered stage often when they're in elementary
school where if you pour water from a squat glass into a tall thin glass
they'll say there's more water in the tall glass later on around 10 to 14 they
go to a symbolic stage where they say oh there can't be more water because where would it go they use reasoning or at
least in Swiss French children of that age is not as not a lot of evidence that
American children ages 10 to 14 ever get into this stage
we built a barrier they are called television
American psychologist Bruner did an
experiment with the kids in the water pouring and after the kid had given the answer there's more water in the tall
glass he put a cardboard covering up the tall glass and asked the kid again into
his amazement the kid changed his answer he said oh no there can't be more Bruner
took the cardboard away the kid changed back Bruner took put the cardboard in
the kid changed once more so if you have any six-year-olds you'd like to torment
so Bruner's idea was that instead of
there being stages of development like going from a caterpillar to a butterfly
that instead we appear to have multiple mentalities you saw two of the
mentalities operating between the visual and the symbolic in the thing I did with
the upside-down faces we have another one that's body centered it's the one
that allows us to touch our nose when we haven't been drinking too much touch our fingers behind our back it in fact is
the one that makes us feel that we are in the world it's far less of a trauma
to be blinded and still feel that you're in the world to know that your body is
here then they have a brain lesion that leaves us limb physically on your body
but subtracts it it's like when you have a stroke people often think of the limbs
that are left as being somebody else's limbs because they can't be theirs they can't feel that they're there and so
each of these ways of knowing the world and these aren't all of them make up a
kind of a patchwork quilt in the way we deal with things and this is why it is so difficult to teach people things is
because often we try and teach through the symbolic and in fact the parts of
the mind that we want to have learn can't understand English we have to
learn different you want to learn physics the way you do it is not by teaching mathematics I think how many
people here know physics or physicist types okay how much of physics
problem-solving is intuitive before you
do the math give me a percentage
yeah yeah I would say it's I would say it's like 75 80 percent for me I
remember when I was in school doing physics and I once asked the graduate
student TA said how do you solve these damn things it says it's easy to just
guess the answer and that will tell you how to do the math
yeah well that was that was the heck that the heck of it was that he was
right because I realized that the next year I took a differential equations
course and I realized immediately the differential equations was entirely
about guessing the answer by using some form of intuition an intuition is likely
to be these other mentalities that don't talk so much but are other ways of
knowing the world and other ways of coming to conclusions what is the import of that when we deal
with technology we have to appeal to what's actually going on in the human
brain we have to take these three mentalities and apply them somehow so
here's what we did when we dealt with the problem of making millions and
millions of computers that were going to sit on people's desks first we have to
put the person into this world in some way the way we do it is by giving them a
way not just a pointing at things but to move things around in it user
friendliness in these graphical user interfaces is not because of the
graphics but because of your ability to pick up things and drag them around it
connects you at a very deep level with this world it makes it less alienating
why do we use icons and images and
windows if I put up pictures of a hundred animals on the wall you could
find the elephant picture four times faster than you could find the word
elephant if the words were place where the pictures are the reason is that
search for the image is done four times faster than you can search for the
symbols the other thing is is that your
visual system is set up to remember a hundred different things roughly in your
environment without you being aware of it and then as a third thing how many
people have cable television here okay
I'm I don't think anybody's record if they're recording we can destroy the
tape but I will admit that I flipped channels anybody
there should be a 12-point program for people who flip channels but I love it
how many people here have flipped channels into a movie they haven't seen
for 20 years how long did it take you to
recognize that you'd seen that movie before few seconds a few seconds right
think about what that means in the context of your entire life 20 years ago
you went to this movie probably with a significant other that you were paying
more attention to you did not know you're going to be tested on this movie 20 years later and yet after seeing
somewhere between 30 and 90 frames of this movie in the middle without any
preparation you can recognize that you see it before think about what that
means in the context of your entire life and in fact our visual system is even
better than that even better than that can anybody imagine why our visual
system should be that good at remembering totally random things
survival why
yeah most of the psychologists nowadays think it has to do with me having to
recognize that you've been in the same place in something like a jungle that's
mostly green okay and what it means is
that you the everything is almost random it's almost the same color and yet you
want to remember that you've been there before so in fact what what seems to happen is that almost every image that
every scene changes that you've ever seen has been encoded in some way so
that you can retrieve them when you see similar ones and compare them it's quite
remarkable it's led to something that McLuhan predicted thirty years ago about
television that as he said television is eventually going to wind up resembling
it's commercials and he predicted that
the shot lengths for instance shot the length of it of a shot in a movie thirty
years ago is about 15 to 20 seconds before they would change the camera
angle or something and he predicted the shot lengths would get shorter and
shorter and of course there are people who have been measuring that every year last year I can tell you that for an
hour of network television the average shot length was two-and-a-half seconds
now compared to the typical primetime
commercial which is one point six eight seconds average shot length in other
words what television is doing is being interesting to us in a visual sense not
having to have any content whatsoever
people will watch snow on television because the visual system is interested
in changing or watch the fire or watch fish swimming around an aquarium because
the visual system is interested in changing scenes and so we take advantage
of that incredible visual memory to offload some of the short-term memory
problems of dealing with computing system and finally you need a symbolic
underpinning we use the language called small talk a long time ago nowadays on
systems like the Mac there's hyper card on the IBM PC there's a system called
plus that's quite good and what these are are kind of a symbolic render
rendering there's a symbolic layer underneath the images and the things that you can move with your hand that
allow you to use symbolic processing and dealing with all of these things the
result is a system that is much easier to use a system that is now taking over
it not just the Apple world but the IBM PC
world as well so what does it mean to take visions into reality
well first you have to have a really good vision and the goodness of the
vision has to be in its persistence not just its details so I invented the
overlapping window idea more than 20 years ago and in order for it not to
have gotten completely barnacled up it had to have something that made it
distinguished from the other ways of thinking about doing those things but
the most important things about the visions are they have to hook up to the human psychology and even harder they
have to hook up to a whole bunch of people with money
who think in the old way so you know
Winston Churchill was an incredible
ad-libber and he was once at a party
the hostess came up to him and said mr. Churchill I just saw a famous Duke steal
some of my silver salt shakers what am I to do Churchill thought for a minute he stuck
his cigar in his mouth he said I know
what to do madam he went over to the earl on the way he took a silver salt
shaker and put it into his pocket when he got over to the Earl he took it out
and he said I think we've been noticed perhaps we should put these back to the
bottom line I'm turning vision into reality is you must make the people
you're dealing with part of the same conspiracy thank you
[Applause]