The Computer Revolution Hasn't Happened Yet (2007)

From Viewpoints Intelligent Archive
Revision as of 22:24, 26 November 2021 by Ohshima (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
thank you for
inviting me to give this talk i wasn't able to
travel to italy this time around
and so we tried to capture the talk as i would
have given it in in rome today
this symposium is about
the evolution of digital media
but i've always been more interested in revolutions and i'd like to
case in this talk that the computer revolution
hasn't happened yet and we've seen some
of the beginnings of it but i'd like to make
the history of printing
and first looking at the printing technology
which happened around 1450
or so using
some ideas that the chinese had done much earlier
but adding to it the
invention of how to make lots
of high quality pieces of type rapidly
and cheaply and so this led to an explosion
of presses across europe
and one of the technical futures for printing
was web printing which
started happening a little over 100 years ago powered
by the industrial revolution the
less expected future is what's
the last 25 years or
where printing via
inexpensive laser printers has moved
into people's homes and reading has
started to move away from using paper at all
all connected by a worldwide
network the publishers
25 years 35 years
ago didn't really understand it and didn't
believe it and now they're they're trying to cope with this
but i would claim that this is not a particularly interesting
future for printing another
looking at what happened to printing is to look at how content
changed so here's a gutenberg bible
big book like
this very heavy looked as
much like a handwritten manuscript as possible
in fact they brought in people who call rubricators
who did the illuminations by hand after
the ex printing was done relatively inexpensively
so they got something that looked like a proper
book to them but
most of the books that have been printed since
then were not like this so the this first
edition of
printing was not like what the future printing
like a printer
in venice my favorite
hero aldous monouchious i
like to think of him as aldo monuccio made
small books he had to invent
more readable fonts to
read things in most
of his books were not about religion but about philosophy
some of them were in greek and
latin but many of them were also in italian
d eldest sent
his sons all over europe to
try to find material to print in these little
books like this that he called the portable library
these books were inexpensive enough so that
they could be lost without it being a
tragedy unless printing started to
spread in a very different way
so that was about 50 years after the
invention of the press about 1500
and 100 years later we start
seeing parts of the real revolution of printing which
is the new ways of thinking and arguing that were
bought brought by the press another one of my heroes
another italian galileo
started looking at the world in a different
way along with kepler and a few others
and science started to get invented
and about 60 years after that
newton
completely revolutionized the way we think about the world
with his principia mathematica which
talked about how
forces masses and
gravity motivated the system
of the world and about a hundred years after
we start seeing governance
changing in a big way for instance here's the constitution
of the united states of america in the late
18th century and so what happened
here was the printing press stayed more or
the same during this period but
the way people started using it and thinking about what it would mean
to be able to present closely
argued reason and
reasonable ways of thinking about the world backed
up by evidence this changed the way we look at
world and one
way to look at this is that if you were born in 10 000
bc with an iq of 500
get very far you might be able
to outwit your fellow cave people
fellow villagers but
iq is not the dominant thing in making progress
humans band together
we have culture we have a way of passing
on solutions for living ingenious
things that we found out about the world we
call that knowledge but we have to ask what kind of knowledge
because this is tempered by the outlook
that we have so for instance
traditional cultures
have an outlook that the world is motivated by
spirits of some kind and the way we deal
with spirits is to pray to them and appease them
engineering
um three four thousand years
bc started bringing up kind of gear-like
chain-like ways of thinking about how cause
ffect articulate and so we have
a kind of a progression of starting with
engineering and the invention of mathematics
greeks and then many
in the 16th century
sort of the first invention of real science by
newton and
in this last century science
has changed and become much more ecological
about it looks at systems rather than
articulating gears and one way to
that if you have a
clock and it doesn't work you can fix it if you have a system
doesn't work you can only negotiate with it
and so these are very very different ways of thinking about
knowledge
and when you have a system you usually the
most interesting relationships are not additive
so if you put a rabbit into australia you
do not get the old australian ecology plus some rabbits
you actually get a completely new ecology
one in this case that was detrimental
to the existing ecology
one of the ways of thinking about changes in systems
is that what we think of as normal
the stuff that we don't notice is qualitatively changed so all of
a sudden we notice it for a while until we get
used to it and many of the most important things
that we think about are ecologies human minds
societies and human media
nd discourse
itself what we're doing right now is a kind
of an ecology and it's driven by
ideas but everything reacts and co-evolves
so that ideas need language to
talk about them rhetoric to express them in various
ways we write things
down in various ways to transmit them to other people
of literature the methods that we use to
deal with the ideas are conditioned
the things that we can learn and
this leads to new ideas and and so forth
one of the most important things is that the notion of language
expanded the greeks did this but galileo
great line he said that the universe
cannot be read until we have learned the language and become
familiar with the characters in which it is written
and he said it is written in mathematical language
and the letters are triangles circles and other geometrical
figures without which means it is humanly impossible
comprehend a single word so what galileo
is talking about here is that the normal
natural language discourse may not be strong enough
or pointed enough to deal with all
of the ideas we have a simple example of that
is just to think of the conventions for making
a modern map of the world there's very little
of natural language there but many many
geometric and graphical conventions
and so when
the computer comes along what we really want to be interested
is what is unusual
about it what what is it doing that is not
just replacing the old media that it's displacing
and this is douglas engelbart in the 60s
the guy who invented the mouse this is a picture
taken last week but it was taken
40 years ago
with the start of the invention of personal
computing and i got interested
in this and thought of doing some of englebart
stuff on a desktop computer
and this is a machine that
i and ed cheadle made in the late
60s and while we're working on this i started getting the
idea about a children's computer
that was the size of a of a notebook and i got this
idea because of work that seymour papert had been
doing with children and for example he found
that if you use the body-centered
young child's view of the world
for thinking about mathematics then you could get
them to do mathematics that they normally wouldn't be able to
do for for many years in the future so for example
we normally think
of a circle as being x squared plus y squared equals
r squared in
analytic geometry but to a child
the circle is making a circle with their body so
they're going a little and turning a little and going a little and turning
little if you make a programming language that uses
that idea like forward five and turn five
and can repeat it over and over again then
you can make this mechanical turtle or in this case a screen
turtle draw a perfect circle
and this is because
this form of mathematics which is only about 150
years old partially invented by
carl gause called differential geometry
is actually in the cognitive space
of the way children can think about the world
so these ideas about personal computing
led to a bunch of us going to the xerox
palo alto research center
35 years ago and starting to invent
various of the technologies that we
use in personal computing today like the
personal computer with bits bitmap screen
and laser printer and the ethernet and so forth
all the while working with children the reason
working with children is because
the printing revolution took about 150 years to really
happen so in the 17th century
not a single person was alive
who had been alive when the printing press
was invented it took normal
humanity regular people
150 years to start
understanding what the printing press was about and this only happened
because children grew up in this technology and gradually
came to understand what it was about now of course
aldus and his friend erasmus
understood this very early
and there were people in the 16th century who understood
this very early but for the large mass of people to
understand it it took much much longer
and so part of the effort here at xerox park was
to try to short circuit this
as quickly as possible because we wanted to see
the printing revolution before all of us died
the last 25 years have commercialized these
inventions so these are all things that were originally
done at xerox park the desktop and laptop
computer ideas the overlapping windows
and icons desktop publishing and wysiwyg
the larger research community
that we're part of and
xerox park together did the internet the laser printer
object-oriented programming and ethernet and so forth so
this is a large payoff
of these ideas done
in the research lab in the 70s but there's
a real problem and that is that the
commercialization of this stuff
left out the most important piece of this
technology which was a computer for
children the computer
for the people who are actually going to bring on the real computer
revolution and so pretty much most of the last
25 years has been just
digitizing old media
so we have digitizations of texts and of pictures
and of movies and of
sound recordings and a few
linkages but pretty much none of
what the computer is really good for has
come into the world of the ordinary computer
user in fact many people think that
what they have today is all a computer can do
and in fact the the opposite is the truth
what the computer is doing today for most people is one percent
of what it's really interested
so a few years ago several of us old
timers got sick of this and decided
to make a computer's children children's
computer happen all by ourselves and this is spearheaded
by my old friend nicholas negripati
and it was based partly on the idea that
a dvd player with a small display
and requires a computer for doing signal
processing can be made to be sold
for about 122 dollars and so we thought
it should be possible to actually make a design that
would be around a hundred dollars that could
be aimed at the
third world initially
and how do you make such a machine well part
of it is to look at where the costs go in the computer that you have
so about half of those costs are sales marketing
profit and distribution so we set up a non-profit
called one laptop per child to do this so
nicholas sells directly to countries
and so we don't have any costs
for sales marketing profit and distribution the countries
do the the distribution
for-profit software like microsoft
and so forth we use open
source software which is free
so there are no costs there
the display on a computer is the next largest
cost item and we invented a new kind of display that's much
much cheaper and the next high cost
item is the hard disk and we use a flash
memory like the ones
that you have in your cameras to cut that down and so
the result here is about 100 dollars
and you turn this into a real
machine you get something that looks like this this
is all the electronics in it not a lot these days
and late last year we started
build our first thousand in
a factory and put some of them into
some of the countries that are going to to buy these machines
so this is about roughly about two years
of work we're part part way into
the the third year
and the machines are full function and they have
a novel network which
automatically links up when any computer is on and it forms
something like the old arpanet that we did in the
60s and if any one of these
machines is within range of an internet
node then all of these machines are connected
to the internet one of the secrets
of the machine is this marvelous display
invented by mary lou jepsen the chief technology
officer and it has very
very high resolution about 200 pixels
to the inch 1200 by 900 and
works very well in this in sunlight and
also with backlighting
one-seventh the power consumption one-third the
so all of these things together are really the
gating technology that allows a really interesting
100 computer to be done and of
course people will use this machine
to imitate the old technology the
high resolution of the screen allows extremely readable
text so here's a page from a book and here's what it looks like
on the machine and this doesn't do it justice because i'm
showing this slide on a 768
by 1024
format
which is much coarser than the screen
which is actually 900 by 1200 here and
then we're making this into a video which
is making it even more coarse so it's worthwhile looking at the screen
in person
and of course just being able to do that for a hundred dollars
is a revolution of a kind
a small revolution because you can make books
and put books on it at the cost
of about 0.2 euros
200 to 500 books or so
can be put on this machine
for the total cost of a hundred dollars
but
he very near future
years this machine
is going to go from a hundred dollar machine down to a fifty maybe
even a twenty dollar machine and this is because
the new technology of conductive
plastics that have semi-conductive
properties can actually make these
not just flexible machine displays but
can incorporate the computer technology in there as
well in a in a form that is much
to make much more robust
much smaller
pretty much all good things and no bad things
we can expect to see over
the next 10 years or so and this will be another
kind of technological revolution
but if we look at the goals of
making a societal amplifier that's kind of like printing
then if we come down
here and the easiest part is to make this hardware
sounds difficult because nobody has done it before
but it's actually partly because none of the manufacturing
interests have been
interested in it so this is pretty
down here free software
we talked about i'm using
to give this talk the actual software that will be
on the 100 laptop so this is a 100
laptop presentation
and um
various of us have been working on content and
pedagogy over the last 35 or 40 years
we know a little bit more about it now than than we used to
and we need to explore this a little bit to
see what kind of thing children might do but
i'll come back to this at the end of the talk
has some new things that children can learn
is that children learn in a social
system that involves mentors of various kinds
and of course a lot of mentors are their peers
who induct them into the what you might call the
popular culture the village culture way of
doing things but parents and teachers are
key and in most parts of the world parents
and teachers don't know a lot about
math and science and so the mentoring problem
is severe in europe
and the united states in japan
and of course it is most severe in the third world
where the average education
of an elementary school teacher is itself
only about sixth grade so this
is a problem that
is a tough one because we can make millions of
100 laptops and distribute them
just in a few months but we couldn't make
a thousand teachers uh in
four or five years so
this gap between the environment
in which learning is done and the
technology that can amplify the learning
is getting larger and larger one of the possible solutions
for this is to make the technology be more
of a teacher i'll talk about that just a little bit
more but right now let's take a look
at a couple of the things that
might be done on this so the first first thing
i'm going to do here is to make
[Music] one of the first things
most children make on this
which is a little car that they can
drive and
so i'm going to paint this car
from the top and
paint it pretty much as a as a
child would do it
putting in a few little
reflections there and most children
to put very powerful large
[Music] wheels on the car
so they wind up making something like this
and so far this is
just a drawing i can move it around
i can make it larger
and smaller i can give it a name i'll call it
but i can also look inside
of it and i can see some
car has so for example
here's a behavior called car forward
what happens if i move the car forward well
it goes forward and what happens if i tell the
car to turn and if i want to do have
the car things do over and over i can just drag these guys out
here and start the
car going and we see we've got now something like seymour pappard's
turtle but with a with a costume
on it and the children can experiment
with steering the car here
by changing the numbers here so what if i
go down to zero
the car goes straight if i go
negative numbers starts turning
but this isn't really like driving
a car what i really want to do is to make make a steering
wheel and so i'm going to
just draw another steering wheel
ike this
and it's exactly the same kind of thing as a car but i'm going to give
slightly different behavior so if i
if i look inside of it let's call it
wheel here as long as i'm here
if i look inside of it
i see it has the very same kind of stuff inside
and this forward and turn
thing called heading if you watch this number
here when i turn the wheel see
that goes positive and negative if you
remember positive and negative numbers in here change the car
uh car's direction and so if
i pick up the name of those numbers coming out of the wheel
and drop it in here then i might be able
to steer the car
by turning the wheel
this is very pleasurable for the children but the children have
also learned what is usually a hard to learn idea
for children which is the idea of a variable
and our experience has been that children learning
a variable this way never forget what it is
and how it how it actually works of course
there are many things that can be done here but
let's take a look at a slightly more serious aspect where
we make a we have a couple of cars here
what we're going to do is make these cars
really small and pretend that they're people
and have them move around in a village and we'll say that
the red one is infected with some sort of disease
and the blue one is not infected
so with just two cars in there we see we have just a couple of
dots here and if i tell them to start moving
you can see the probability that they're going to
touch each other is going to be very very low
so this is a disease that is
not very communicable so
i'm going to stop that now what happens if we have
a thousand
villagers and here's the red guy the one
infected now we can see that they're very
together and so the probability that this red guy is going
to be high
and when he touches the blue guy will make him red also
let him be infected also so let's see what that looks like
so it happens very quickly and here's
the all of the ones getting infected and
if this is a deadly disease then they've all died
here happens very very quickly
and we can do this very easily on the computer because
the computer can make things very
very quickly and inexpensively so let's
try this with 500
the probability is of infection is a little bit lower
but we can see as soon as a few get infected
the disease here is very efficient and
all wind up dying
so what if we go to let's say 80.
so now it's quite sparse
hardly anything is happening
but
in fact
because this disease is
contagious and has no cure
eventually they all
contract the disease and they all die so
ne of the problems in the world today is that we have
some diseases that are like this last one
which take a long time to even be noticed
so this first one i showed is like typhoid
happens very rapidly it's very dramatic
people start doing something about it but this last
disease is like aids aids takes
about five years to incubate
and so if you know that diseases
contagious deadly and incurable
are always going to wind up
doing this then you might take
uh action much
much earlier even when it doesn't seem like
is going wrong if you don't understand this about diseases
then you might use your common sense
down here and say oh nothing is happening
therefore we're not in danger let's just ignore it
now the sad thing is is that
in the underdeveloped parts of the world about 100 million
people are now infected with aids precisely
because the common sense reasoning
was not strong enough to deal
of of threat
whereas a child can very
easily sit down and experiment with a
little script experiment with
contagion and quickly
understand in a visceral way
when
something is detected that is deadly
even if it takes a long time that something should
right away and by the way
the the other little known fact in most parts of
the world is that the largest factor
in dealing with infectious diseases is actually
sanitation it's you know
we have wonder drugs and and so forth and
eventually these are likely to
be stronger than sanitation but right now the strongest thing that we
have is sanitation the understanding
ability to imagine very very tiny
organisms that are trying to eat us
all of these things that are very difficult to think about
can lead to actions in
the large that will save hundreds of millions
o another interesting area
is this form of thinking that's called mathematics
and it's generally thought
that children can't do this very well and so we teach them
simple arithmetic and
we wait until high school to try and teach the mathematics and this doesn't
work very well the truth is the children can actually
learn mathematics relatively
easily if you put it into their context
and one of the places in the world that's most
famous for doing this with young children are
the fabulous kindergartens at reggio emilia
which i've been to a number of times and they are truly
inspiring out of kindergarten in
the united states first grade we
found a marvelous teacher julia nishijima
who was showing her
children how to do real math also so
in this case the idea is you take a single
tile of some kind like a square or a
triangle or a trapezoid or a diamond
and make progressively the next larger
sizes of that particular shape
you can see trapezoids here a little bit
challenging but the children do this
individually and then
one of the things that she did was to after the children had done
this constructive exercise she would get them to
try and look at it in a reflective way
thinking about it much slower or less action
and try and write down something about what they've done
so in this case
lauren here noticed that the first one took
one and the next one took three additional ones
the next one took five additional ones the
next one took seven additional ones and the total
was one the total was four total
was nine and quickly she saw oh these are just
the odd numbers and the
total numbers are look like the square numbers you can see she wasn't
oo sure about 6 times 6 here
but she discovered two progressions
here that have great interest beyond just
growth laws these two progressions are sometimes
called first and second order differential equations
in this case they're discrete the
uh constant uh that
drives this one is two so there's a two
between each one of these and then if you add
the the two of these to this you get
and if you add four and five you
get nine you add nine and seven you get sixteen and so forth and
so this is a generative scheme for growth but it
also happens to obtain in much
elementary physics of the kind that newton
invented and
we can for instance look at this
terms of making our toy
car here move across the screen okay so
one of the one of the things that we can do here is to
think about using these progressions for a
ride wide variety of phenomena like
speed and motion so here we're going to set
the car's speed to 30 each time
and we're going to change
the car's position on the screen each time
further to the right by
that car's speed so let's see what happens we'll drop a little
dot here and we see
that these dots are evenly spaced 30 apart
and the car is going across the speed at a uniform
rate
so if we make the
car's speed 50 here
the car goes faster but still at a uniform
rate now
what if we instead of making
the car's speed constant we will actually
reset it
and get the car's speed to increase
by 20 each time and go across
o this
is what the kids were measuring with
their little little blocks except the cars
the speed of the increase in size
was two and you get this
growth law of the speed increasing
constant amount and the distance increasing by an
exponential amount and of course
for races
what we want to do is to do something more
like use a random speed so if i
go in here and pick up a random
tile here and say make something
like 50 and
now the car's speed will be
different amounts we can see the car
going jerkily forward and so forth so this is a way of modeling
uh speed and
acceleration in a way that children can understand
now let's visit another
favorite italian here galileo
who is very interested in speed and acceleration
and particularly about falling objects and
let's listen to this classroom of 11
about how fast or how
slowly objects of different weights
might might go so i'll play this little movie
and the objects that you think will fall to the earth
of time
they think the heavier one is going to fall faster
both hands oh do not
how do you determine this
o we decided let's
drop them off build them
at the school and they'll use
of course the stopwatches don't
work very well when did he actually drop it
[Music] i think we should
do the shot put and the sponge ball
because they're two totally different weights
and if you drop them at the same time
maybe they'll drop at the same speed drop both of them
together which is exactly what galileo did
we get a direct
way of looking at this so important thing to think about
is that aristotle didn't think to do this
and st thomas aquinas didn't think to do this because he believed
in aristotle and galileo
is perhaps and we know that probably
aristotle didn't ask a child and st thomas aquinas
didn't ask a child galileo might have been the first
adult in thousands of years to be able to
still think like a child we found that
in every classroom of 30 children there's usually one child
who will think that you should just drop the
see what happens
and so we get this
qualitative result that
things are falling somehow
the same way regardless of their weight well
actually we can go further than that we can
take a movie of how
things fall and
but how can we understand it so one of the things we can do is single
movie but still pretty difficult
in this system the movies
are actually made out of
the same stuff that the children manipulate and so
we can pull the frames
out and we can even stack
the frames up and when the children see this stacked
pattern like this they immediately say accelerations
and so what kind of
acceleration and what they do is
they measure from the bottom of the ball here to
the bottom of where the ball has gotten to
about five frames later and
when they measure from the bottom of
this ball to the bottom
of where the ball has gotten to five frames later they see that this
a bigger than that one of course that's what we see here
so some acceleration is taking place
and so if we measure again
here and measure again
we can compare these by
stacking them up let's see if
enough if i stack
it up here
like this
and we see that this little patch here
is the amount more that the ball
traveled the next time and the patch here is the amount
more that traveled from here and the patch here is the amount
more and so this is the acceleration
of the ball and looks pretty constant that's
what the children were playing around with their cars so they very
quickly write a program
that's just like the one that they did for the car where the
they make a simulated ball here and the ball speed
increases by some constant amount and we
change the vertical position by the ball speed and
so if we started going like that it
seems to fall but how can we
determine that
the what we've got here is
an accurate model of what was happening in the physical world
and here's how tyrone decided to do it
and to make sure that i was doing
it just right i got a magnifier
which would help me
figure out if the size was just right
after i'd done that i would
go and click on the little basic category
button and then a little
menu would pop up and one of the categories
would be geometry so i click on that
and here it has many things that
have to do with the size and shape
of the rectangle so
i would see what the height is i kept
going along the process until i had them all
height i subtracted the smaller
from the big one to see if there was a
kind of pattern anywhere that
could help me and my best worked
so in order to show that it was working i decided
to leave a dot copy so that it would show
that the ball was going at the exact right speed and acceleration
so we can see that
tyrone decided to just leave dots
like he had with the car and that the dots lined up
with the actual physical phenomena and that
is a way of showing that this is was an accurate
way of modeling
the constant acceleration of gravity
the the
xponential increase
in the distance covered another way to
do this is to just run the simulated
ball against the movie so we can see here
that they're tracking along very well
and that's another way of verifying that the
mathematical model here is
actually uh in
accord with uh nature and
of course there's a nice little payoff once
you've done that once you've made a model of gravity you can
apply it to anything so why not make a spaceship
trying to land on the moon and if we start
it going and don't do anything then this spaceship will just
crash and
if we put a motor if we think of
gravity as a velocity eater a motor
is a velocity producer
so that's a little one line thing that connects up the
the joystick
here so i'll start this off and
turn on the rocket motor here
and the effects
of showing the flame here is this little script
and the crash if you're going too fast is this little script
this is a game that people used to pay money for called lunar
lander but it's within in the the range
of an 11 year old or a 12 year old
this is so the computer allows
to experience the thing that
mathematicians and scientists
know about math and science and that is it's fun
the the reason we do it is that it's fun
it's illuminating it's exciting it's
art and it's fun
and i don't believe it's possible to learn math and science without
being introduced to the fun
the intrinsic reasons for doing this stuff
and so we see that the computer has something very very special
that books don't have
because computer in a sense is a book that can read
itself it's a book that can go
from claims in language
to being able to carry out these claims in simulations
this gives us a different way of both learning
about what an epidemic is but also arguing about what an
epidemic is learning about what
physical dynamics are arguing about what
physical dynamics are and it's a way of
doing this for children at a much earlier age
than has ever been possible before
and now on the 100 laptop it can be done
really inexpensively and be sent around
world so to
finish up here i just want
to the cautionary part of this is that
just like the potential for democracy
was in the printing press in 1450
but it took 200 years
before people were even strongly writing essays
about it and it took until the 18th century until
it really started happening the potential
for science
was in the printing press in 1450
but it took again
until the 17th century before
people a few people started looking at the world
in a different way started writing about it got a few more people looking
at the world in a different way
so the big revolutions
are a bit like epidemics
there are ecological phenomena and they take
a long long time if the revolution is going to be a big one
most people don't notice the revolution happening
at all and certainly with computing most
don't notice the revolution
at all and it could take
another 50 years unless we can do
something about the mentoring systems so i
urge all people who are interested in this problem
to think about mentoring think
about ways of changing
the dynamics of mentoring in our society
both through humans on the internet
better attention to teacher training but also
to thinking about how the computer might be a
better teacher than it is today thank you very much and thanks