The Dynabook -- Past, Present and Future (1986)

From Viewpoints Intelligent Archive
Revision as of 19:15, 5 December 2017 by Ohshima (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
well the Greeks held that the visual arts were the imitation of life but the
computer arts are the imitation of creation itself thank you very much
[Music] [Applause] [Music]
my name is John Jacques I have both a great privilege and the challenge of
introducing LNK our banquet speaker why
why do I say that it's both a privilege and a challenge well my first reaction
when I was called up and invited to come to make some brief remarks and introduce Alan was that I was very flattered and I
really appreciated the invitation that although I'm in the venture capital business these days here in Palo Alto I
spent 14 years at Xerox and I was in fact hired into the Palo Alto Research
Center shortly after it was formed by Alan Kay and one of the many anecdotes
I've chosen not to tell tonight is about exactly the manner in which Alan hired
me which is a great story that I'll save for later when Alan and I decide to
embarrass both of us or at least when he will tell his version and I'll tell mine
but it was a wonderful time working for Alan and many of you of course know him
and it was a period where I got to participate in some tremendously challenging and wonderful research and
we have lots of wonderful stories so I instantly accepted foolishly and and as
soon as I hung up and I thought about it for another 30 milliseconds the
trepidation set in because how do you go about introducing someone who needs no
introduction to use a trite phrase and as I said I considered lots of stories
and anecdotes and I concluded that I would just stick to a few of the
formalities and then just a brief observation as I said many of you many
of you know Alan of course and and and I'd like to think he needs no introduction yet conversely I'd like to
think that there are at least some people here who were not only the participants back at those times and
that we're not just talking to ourselves but there might be a few people who don't know that Alan of course has a PhD
from Utah he taught at Stanford join Xerox PARC went on a sabbatical to USCIS
I went from there to Atari and is now at Apple as an Apple fellow I think that's
about right and that's all I'm gonna do on that Ellen those credentials are not the reasons of
course why Alan has been invited to come here tonight rather it is what I would
characterize as almost two decades of a commitment to an idea that spans more
than the individual contributions and that is the idea of powerful personal
tools including but not limited to computers for children of all ages
Thomas Kuhn in his book the structure of scientific revolutions
Revolutions talks about shifts in paradigm and that those are the important inflection points in the
evolution of science and I believe that's what we've seen with work
personal workstations that they represented a different kind of paradigm
for computing and when you step back and try to integrate over all the individual
contributions in many of the speeches we've seen here the work that I and others were involved in I really think
it is that shift in paradigm now now having that idea of that alternate
paradigm of personal workstations and powerful tools is necessary but it's not
sufficient because in order to succeed you have to be able to lead others to
absorb them to accomplish more than they thought possible than they thought that
they were able to accomplish so you have to be able to establish a vision and then describe that vision and Alan has
been able to do that to create those targets those images in both words and
pictures many of you are familiar with some of his expressions the pursuit of
the Holy Grail the establishment of a target that we can then try to reach even if it appears to be well beyond our
reach the Dynabook of course is the most well known of those although that's only
one of many and those of you who have had a chance to dig through some of the papers will recognize other visions and
symbols that were used including the reading machine the kiddy camp the mini-com the flexing machine also known
as the reactive engine and probably many more that I don't know about that one in
addition he combined that vision with the ability to express it in ideas and
in expressions that we will probably not likely forget many of us have heard the
exhortation that software is of course just like a waffle you have to be
prepared to throw out the first one we all took we all took part in a in a
medium that and in an off-site session at fahara dunes which many of us would
have thought was a retreat to go talk about strategy which Alan rechristened as a strategic as a strategy advance and
at that meeting he exhorted all of us to break out of what we had done and to go
back and burn the disc packs on the
other hand not all of those expressions were associated with the computer business and one of the fondest memories
I have which I was thought of as I drove down here today was one year many years ago Alan I think we went out for a drink
with Bob Barton who had come to visit and we went to a local watering hole in
Palo Alto which is now closed and turned into a Mexican restaurant and as we
walked in and we're going to stop and chat a little bit we walked in and Alan sort of looked around and in his
innumerable style look around so yes a nice place sort of a plush sewer and
he's had a tremendous ability at creating images that can motivate us
establishing targets and without further ado and without further comment let me
invite Alan Kay up to come and talk about his vision of personal workstations thank you all very much
[Applause] [Music] [Applause]
now I was reading Toynbee the other day and I know he's out of fashion but there
was a section that was particularly interesting which is how do you judge a civilization do you judge it on its art
its music its social functions its
government I think 100 years from today our civilization will be judged by the
complexity of audio-visual equipment necessary necessary to put on a
presentation like this and I think we
should all look at one gloomy note before I start and that is to realize
that there's not a single computer involved up here helping a show anything
here and it also at this point like to ask everybody to raise their personal
computer in the air can I see it I rest
my case ok let's can I start out for the first
slide please now I have to warn you that the to me
the coupling of work and station is the ultimate oxymoron because I titled my
paper you have to be finished playstations because because most of the
stuff that that i was interested in had to do with advancing forms of play and
when I was a kid I got very involved
with Walt Disney's Fantasia and these images like these that entranced me now
my background is my father was a physiologist and my mother was an artist
and a musician and I grew up working both sides of the street going along and
Fantasia completely blew me away as a sort of a way of combining many of the
senses together and I thought gee it would be really neat to be able to do
your own Fantasia and I thought nothing more of it for for puberty he had
happened and one does not think of thoughts like that at age 14
and of course that one is is the kind of
thing you think of at age 14 but you
know we just have to admit there is a little power urge in all of us and that
image particularly struck me now also at
when I was a teenager I started reading science fiction and I discovered that
Robert Heinlein would occasionally mention things in his science fiction
books that actually had existed and he used them as context and one of the things he mentioned was the two Kista
scope which I discovered actually had been invented and was used for measuring
reaction times and ability of people to see whether they could read a word and
500th of a second the other thing that Heinlein mentioned was Memex
which he used as a generic term and I had learned that if I went to the
library and poked around long enough I could find almost any of hymens references and sure enough I found as we
may think written in July 1945 by Vannevar Bush and I read it with great
interest and I think that I would like
to sort of revolve my talk around bush in a couple of ways because I think the
the power of Bush's vision was not that
he tried to extrapolate from technology but rather he took what should happen
and then tried to justify it by looking at technology that could could happen in
other words he took a proactive view rather than a reactive one and my belief
is that in fact Bush would have thought up the Dynabook if he had his work
habits involved going to all-night restaurants
but in fact he was a creature of his times and like to work at a desk and so
his envision of what a personal workstation should be was a desk and
when I started thinking similar thoughts I found long before that I could never
work at a desk and in order to get anything done I had to go to an all-night restaurant and take showers so
some of the early ideas for the Dynabook was that it'd be very very portable and
waterproof now I guess the other thing I
should say before I launch into this thing is that the all the stories that
we've heard at this conference are success stories except mine and I
consider it my job my duty to fail as
strongly as I possibly can in a direction because I think that the one
of the differences between practical research and sort of far out research is
the difference between goals and directions in other words we're taught
that it's a very good idea to make up goals and in fact it is compared to just
probing around randomly but I think that one of the lessons of Parc that turned
out to be very powerful was that the most powerful thing you can do is set up
a set of directions and I think of those directions as being like a magnetic
field I thought of us all as being iron filings trying to line up in a magnetic
field that we're almost all so simultaneously try to create or salmon
swimming upstream in other words far out research is when you don't know what it
is that you're trying to get is done by smelling your way along you're looking
for gradients and so for and you have directions that's good because the goals
are the things that drop off first the goals are the things you succeed at
they're the things you pat yourself on the back but in fact the goals all the
goals that we did at Parc I think we realize are obsolete and yet the
directions we had there are still as valid as they ever were now this is
AGA's ork and if you don't know what is gus' orc
is this is a the console of a 6600 and this is the kind of machine i used to
program in the early 60s and in fact
this machine was a wonderful machine it had it had about three things that would
drive you and absolutely crazy about it but aside from that which it was
incredibly simple and incredibly fast and sitting down at the console of one
of these babies was just about the right
one I was thinking of Gordon Bell today saying you buy computers by the pound
and I think one of the ways of summing up what I wanted was about fourteen
ounces of one of those sliced to order
now in in this first five years of
programming around I saw a couple of great ideas quite a number of times
before I started understanding them at all and when I was trying to write the
paper I was thinking boy it was really amazing to me how many times for
instance I saw object-oriented programming invented by somebody else
and liked it and yet it never hit me for
those first five years let me give you an example here see if that goes off good well let me
just show you something that I ran into
when I started programming in the in the Air Force in 1961 back then at least in
the in spite of Doug Ross is early operating systems the Air Force hadn't
learned about them and Air Training Command had a bunch of burrows 220s
which were modest sized machines by today's standards but in fact they were
rather large physically and one of the problems the Air Training Command had was passing tapes around because the
operating system wasn't guaranteed to run and they had different different
formats and stuff and by the time I got there as a fledgling programmer somebody
had thought of this idea and this is the way they did it here's what a file
looked like on the Burroughs 220 in Air Training Command three sections to it
first section were a bunch of relative pointers second section were a bunch
burro's 220 machine code the third second section we're a bunch of Records
and the way you use this was you read these guys in decor and jumped relative
through these guys which you knew what they meant and that caused the execution
of these guys which you didn't have to know what they were and those guys could get into this stuff which you didn't
know at all in fact didn't want to if that is an object-oriented program I
don't know what it is you're here so somebody some nameless person used this
technique very very early and the macro assemblers of the day back then were set
up to use it and I used it to and yet it didn't hit me how powerful it really was
here's another one I learned real early and liked but still didn't realize the
Burroughs be 5000 that happened to be
the replacement computer for the Burroughs 220 but of course Burroughs
never had a difficult time building this machine but in fact Brahe's be 5,000 had
many many innovations and the one of the
chief ones was this notion of trying to build an environment for a higher-level
language but the thing that it was hard to for me to understand and hard to
explain to other people is that the Burroughs be 5000 was certainly the
first piece of hardware I think that ever tried to make store into a
procedure it did it the wrong way but in fact it was something that was
hardly being thought about back then except by a couple of people like Doug
Ross now when I as a as a programmer back then I had no knowledge of what was
going on in in our core anything else we
didn't you know we enlisted men weren't expected to read in fact we didn't just
like the students of today
so in fact that wasn't that was another time of seeing a tremendous idea then
through a series of misadventures I wound up at the University of Utah and I
discovered a number of interesting things one was that Dave Evans was much
younger than he looked and much older
than he appeared but in fact he is one
of these guys who at age 30 just put his appearance on hold and he's still out
there running running marathons one of the things that Dave liked was Ivan Sutherlands sketchpad
he had a big pile of these brown Lincoln labs documents on his desk and before
you found your desk he would give you one of these and you had to read it and understand it and I'd like to show you
the first sequence from videotape look this is the earliest known movie of
sketchpad this was done in the summer of 62 I show this a lot today
so here's Ivan with the blood starting to run out of his hand but that's okay
now notice he did something that we all are familiar with but now he's pointing
to all these edges and he's telling sketchpad to make them all mutually
perpendicular and sketchpad figures out how to do it and the reason is that
sketchpad is not just one of the first graphic systems but it was the first non
procedural programming system that might
be disputed but let me go on with it here again he's making a couple lines
he's saying make them parallel and perpendicular and sketchpad straightens
them up now he's using a constraint called collinearity and the little
dashes are aligning themselves right
over the guidelines there
okay and now he's telling the guidelines to be invisible and he's made his hole
through the flange you notice that sketchpad is the first system to have a
window as far as I know in fact he had drew on a very large virtual canvas now
he wants to make a rivet and once again he only has to indicate the general
topology and the rules that will constrain the topology into being what
he wants that's going to be the center of an arc and I notice this clever thing
that if he makes these mutually perpendicular that will force the well
you see what I mean there it is and now
if he distorts it it will still obey
that general geometry now he could have constrained the ratios of the sides just
as well and kept the the the rivet a
specific shape in this case what he's done is constrained the angle
relationships and sender relationships so we always get a symmetric rivet now
sketchpad is also I believe the first true object-oriented programming system
because what we see here is actually a dynamic instance of that master drawing
that was just done he called the masters now nowadays we call them classes
because of Simula so they notice he can take that instance and twirl it around
and make it different sizes and he's
going to stick it into the the hole of the in the flange here and one of the
reasons we can tell how early this is is because in this whole film there's no
illustration of visible constraints that cain't was the second phase of the
implementation of sketchpad here's a few more now you might ask why is the screen
blinking well the tx2 here is putting up every dot by brute force so he's gone back to
the master and he says well I really don't want to see those guidelines I'll make them
invisible and notice all of the instances feel them dynamically and
anything that you can construct in sketchpad can itself be a master so this
he's taken these two building blocks the flange and the rivet made it into a
master and now he can create instances from it
well this isn't home cooking I don't know what is this this is the kind of
thing if somebody did that today we would say my god that's unbelievable okay stop the stop the tape please
because one of the a British British art
historian by the name of letter B said that a man a work of art is one man wide
and many man deep and something like
sketchpad has the interesting characteristic is that it's not just a
work of art it's a masterpiece so it is one big man wide and one big
man deep it's most of those ideas were ones that all though I even sort of
gathered them up the data structures came from Doug Ross's papers and there
were ideas here and there and he scrounged the matrix multiplication
program from lesson Ernest and so forth the vision what here was so pure that it
stands up today if somebody were to do a system as good as sketchpad today we
would like it it also is the only known
computer graphics thesis in which every illustration was done by the thesis
program think of that okay so that was
pretty exciting and the other another
thing it was exciting at Utah is Doug Engelbart who in those days used to
travel with a special this is this is
how much he wanted to communicate with people he used to travel with a special 16 millimeter movie projector that could
stop train and he would be there with he would be there with one of these things
and his stop frame thing and the reason is is it's very hard to explain a system
as good as nls in a dynamic movie it's hard to see where the cursor is and it's
hard to see what's going on and so he had he had evolved this style of talking
about it that was really incredible and as I recall the movie he showed then had
a green green characters on a screen and I think it was the 3100 implementation
of the system well it completely blew our minds at Utah
we just could hardly believe what he was
trying to to do and of course it has some roots in Bush's ideas also but in
fact one way of characterizing Bush is that he was trying to be to find
something that was complementary a complement to human abilities this is
the way he writes in his paper and Doug was trying to augment human abilities
and if if I could characterize what I
tried to do after seeing his stuff it was to try and allow the system itself
to be augmented by the human so the augmentation would actually be mutual
back and forth and I still believe that is the main problem of personal
computing but anyway I think all of us
who were at the 1968 fall joined computer conference and I know Butler
mentions it in his paper this was I can't take the time to tell you about it
but believe me it was like magic ARPA spared no expense thanks to Bob Taylor I
hesitate even now to mention the figure that they spent to put that on I'll let
Taylor decide to say if he ever wants to but believe me the the whole objective
of this thing was they had to illustrate what the concept was and they spared
nothing to make it to make it work and
it was an unbelievable thing even for those of us who are fairly familiar with
the system just never forget it ever
yeah I I do remember the number and yes
now there are some other things that aren't easy to talk about but I just
have to say that there have been many people who have had an enormous
influence on me in my somewhat strange
career one of the greatest as is Butler
Lampson and the STS 940 was one of the
things that machine was so wonderfully designed I can't tell you but that was
the machine that Doug Engelbart's stuff was implemented on it was one of the
major machines thanks to Bob Taylor again that we used in the ARPA community
and again it wasn't just the the fact
that it was rather inexpensive and rather responsive and so forth there was
a whole set of secondary things in that system that were so important for
instance Butler had this odd notion that machines occasionally crashed and in
fact the system was designed mainly for crash recovery and so those of us who
used the machine can remember many years in a very flaky era where we'd hardly
lose more than a couple of characters a year if that through many many crashes I
can't explain all the details why but it was beautiful and it made a tremendous impression now
those influences thanks to Dave Evans who hook me up with a crazy Texan by the
name of Ed cheeto resulted in this design in this machine called the Flex
machine this is sort of in a general period 1967 to 69 and while I was doing
this I looked at the the link of West Clark but and liked it very much but in
fact I didn't understand it I originally poo-pooed it because it didn't exit
you to higher-level language and it was small in all of these things and but
I'll return to that story in a minute so here's what the Flex machine looked like
and it's as drawn on its own display and
we actually we like Doug's idea of tying a brick to a pencil as
sort of a de augmentation concept but we sort of thought that's what the mouse looked like - and thought that tablets
might be better which I still think they are so we built this build a table and
windowing was in the air then there was
a lot of hookup between Utah and Harvard where our eivin was at that time wearing
about ivan especially was wearing about 3d clipping and windowing and Cheadle
and I were wearing about 2d clipping and windowing in the Flex machine had a 2d
windows and in fact multiple windows but
of course in a calligraphic display it is somewhat a low-calorie exercise to
overlap windows it just results in a jumbo what ran on this machine was a
simulation language that I had derived from Simula I essentially had an
enormous revelation when I learned Simula the hard way because I had to and
discover that simular was a programming language that created the same objects
that sketchpad did all these wonderful things with and that was an amazing
thing because it meant that in some sense we didn't have to solve the
constraint problem in order to do many of the same kinds of things that sketchpad was able to do with great
difficulty so here's an example of what
the structures in a a flex program of
the day there was a scheduling queue which is very similar to Simula a an
instance that had the local variables for a particular patient the instance
pointed to what we now call a class which contain genetic patient behavior
and that was how the Flex machine was organized this is what the flex code
looked like on the Flex display
this is the flex machine itself was
organized like the B 5000 that was one of the worst mistakes I ever made and
not because the B 5000 was a good idea but in those days I didn't understand a
very important truth and that is that good ideas don't often scale and in fact
we were trying to build a machine that would sit on a desk and we've spent all
of this time worrying about the processor and forgot totally that a
computer is mostly memory so that's too
bad but that's that the machine basically had three virtual processors
all implemented by one set of micro code now reason I'm telling this story of
course is the Flex machine was a complete failure in a variety of ways
technically I think it was pretty much of a success Cheadle was a genius and
did a great job and my problem is when I
tried this system out there was designed for lawyers and doctors and so forth
it absolutely appalled them they shrank away from it like somebody does from the
plague or the way people nowadays do from APL had made it many of the same
user-friendly appearances to it that was the first time ever in my life I
realized that graduate students in the rest of the world were two different breeds we loved it we thought it was
very powerful and oh one final thing is
the Flex machine this is not a packet switching Network but it did use coax
this in fact was what is called an urgency scheduling coax in which the
trans the transceivers themselves have
an urgency and a priority register and they sort of vote among themselves as to
who's going to who's going to get the thing well while brooding about all this stuff
I ran across the work of Seymour Papert and had a revelation and that is boy we
certainly don't wait until kids are in graduate school to teach them how to read or how to use a pencil in fact we
like to get them at things as soon as possible and really in many ways that
the idea that Packard it was worthwhile teaching kids the program and important so forth changed
my whole idea about what a personal computer actually should be and that is
there should never be a thing called a personal computer that doesn't isn't accessible to children there very few
think it's not dangerous like a car guys there's no reason waiting until they're
18 that give them 6 months of driver's ed like the Boyer report says in order
for it to really be a full-fledged medium it has to do two things it has to
be accessible to children and it has to be used for mundane tasks so one of the
things we thought of got early when we were thinking about it is it's a
personal computer if you are willing to
put your grocery list on it carry it into the growth of the grocery store and
out again with two bags of groceries how many people would do that with their
what they call their personal computer today so this notion that it somehow
should fit humanity and everything which is a rather novel one to me was very
important he taught a course called advanced systems design and the first
day he told us well there are a few things known about systems design but they most of them are written down and I
want you to read all of them but my job in this classroom is to firmly disabuse
you of any finally held notions you might have brought here so in fact what
he did was the garbage collect our confidence
that was the best thing we ever had those of us who lasted out the class and
the flying chalk and all app walked out of there feeling that nothing was sacred
including Bob Barton it's very important
and Bob got me reading McLuhan and I'm
not going to try and explain McLuhan's ideas here but the I think the most
important thing that has been said in this century is can be set in two ways
one is McLuhan said I don't know who discovered water but it wasn't a fish
that's the simple way of putting it and the other way of looking at it is in
order to pick out a message from the noise you have to know what the carrier
of the message is so you have to know the difference between a fly speck and a period that means you have to know what
unadorned paper looks like and because you have to learn what unadorned paper
or unadorned television or unadorned computers are like in order to pick the
messages out of them you have to absorb what they are so you become what the
medium is what the material is in order to use it as a tool that's pretty
heavy-duty stuff and because we become it we don't know that we are it so this
idea of actually being able to shape the way people think about things by designing the material carefully was
something that became very important one of the things that of course my old
mentor Dave Evans has been around a little bit he was Butler's thesis advisor my thesis advisor and was on the
alcohol committee and a few other things and one of things that Dave liked to do with his graduate students is to invite
them into high-level negotiation meetings because he figured that
graduate school should be a transient place he hated the idea of it and he
wanted us to learn how to negotiate so one of the things he did is he invited all of us graduate students into one of
the ARPA contractors meetings that was held at Utah and we were there to watch
the big boys deal it out and was it was most interesting I can tell you and at
the end of the thing Taylor asked all of us whether he had any suggestions and of
course there were several and Warnock piped up saying you know if you guys are really
well he said it a little more subtly this and he said if you guys are really thinking about you know we are going to
be have our pH D's pretty soon all the graduate students should be meeting together right now and Taylor thought
that was a great idea so they set up a a similar graduate student contractors
meeting every summer and Warnock and I were the first two to go to from Utah
and I talked about the Flex machine there and Warnock showed his great algorithm but the great thing we saw was
at the University of Illinois which is the very first plasma panel the first
flat screen display now wasn't this large this this display screen violates
Kay's law which is that all one inch flatscreen displays work and no five
inch flatscreen displays work now that
law is still true but in fact they did
have a one inch flat screen display there and a lot of us graduate students
who were there spent the rest of the time thinking about what it would be like to put a flex machine on the back
of that display because if you're doing input an output with a display then all
that cubic volume there is absolutely irrelevant and my new version of
Engelbart's brick on the pencil was imagine dragging around a digital watch
with a CRT on it wouldn't that be ridiculous so this idea
wasn't called the Dynabook then but this idea that this is one of the destinies
for the computer that goes with you started around the summer of 68 and of
course the the idea was that should be down at the lowest level so the kids
this age could use it and this quaint
idea of portability now point five herniation spur block is not portability
so there's this notion that you should be able to carry something else too and
now the link came forward really strongly because everything I started
thinking about then reminded me of the link because the one of the things as you'll learn tomorrow is the link was
designed as cleanly as anything as ever designed it was beautiful you just like
you know it's just I salivating right now thinking of it and that was the first time I learned about it because I
really needed to learn about as to what they were actually able to do with almost nothing it was fabulous and it's
worthwhile studying today let's show the
the second segment now what this system
is is system called Grail done it ran
corporation here's an early menu and
this is the system this has sound on it
[Music] so it recognizes the box and makes one
now it's recognizing as printing here the printing in the box is being used as
commentary only in this case the box is slightly too large so we may change its
size that's where Mack window control came from then drawing flow 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 floor
arrows attached to the process post new area and move the box to a new position
this allows us to draw a new box then
chop off its corner and label it sub scan with the residual error now notice
it misses the end here and he sees it but he can keep on going now he goes
changes is it correct the labels add a
decision on sub scan so that control may float to the connector a zero they can
complete the diagram from scan to post
okay stop the tape okay now that system blew my mind I came
along just at the right time because this system was tactile it was like sinking your arms right through the
glass of the display and touching what you wanted to do and it didn't matter whether they implemented flowcharts or
not but in fact I found I read licks 1960 paper just recently and discovered
there was a section in describing a system just like that and I think the rand people had obviously read it also
but this system removed the barriers you
were right there and mentioned a couple of other things about it you could dive
through any of those flowchart symbols it was really a hypertext kind of thing
a multi-dimensional thing you could simulate the diagrams it
would flash the diagrams there's a speed control up there it was modeless and modeless doesn't
just mean generic modeless means that
you never have to terminate what you're doing in order to do something else and you just saw it there it's one of the
most important ideas in user interaction and this is the first system to really
have it in spades there been other ideas before this is great this system barely
worked was on such an incredible Cluj of hardware and software that eventually
collapsed under its own weight I used to show this movie about every three months
at Parc we used to show you know because if we wanted it we didn't want assistant
had flowcharts but we wanted to be like that we wanted to have that level of tat
we wanted it to be not even visual we wanted to be tactile just had to be
organic you know it had to be messy and stuff okay let's look at the next slide
and now I'm gonna go quickly because my goal is actually to to get to the park
movie in one minute which I'm not going to be able to do so here's just some now
of course I knew we couldn't do a dining book right then but this is 1970 and I
started consulting for Taylor at Park and thought boy we should be able to say take some television sets and this is
the old flex rubber bed sheet tablet if
you want to make one for about 25 cents this is how to do it and using Lisp
techniques now for a simulation to implement the kernel of a simulation
language look like a really great idea because you get an incredibly small
kernel and the rest of the system is self describing okay now what I what I
want to do is degress for just a couple of minutes if I can this is talk is
really in three parts I'm right at the end of the first part middle part is
shorter and what it is is it's kind of a
lie in that this is not exactly the way
I was thinking back then this is after 15 years of trying to explain these
ideas to people but this is what what I'm gonna tell you next is what I think
really drove what we're trying to do and it's a much better story than we were
able to tell to anybody back then this
is Park this is the time to give the
tribute to Bob Taylor he actually deserves another one so I'll say a few more words now you know let's see if
there's one before that okay this is a typical way Park was organized was
organized this way because Bob I think he had two goals and why mine one is he
wanted to be different he wanted to be comfortable and I think he'd also
discovered that it was impossible to leap to your feet to denounce anybody when you were sitting at a beanbag chair
[Applause] there is this there is this curious
passivating effect that they had as you slunk in and I I think there are many
people in the room that remember what they felt like after trying to sleep on one of these because you your body
gradually as you slept sank into it and assumed the strangest positions that you
woke up but in fact this was the way Park was set up I I'm not sure who that
person is there but because and that
youngster I don't recognize either but he doesn't look very trustworthy to me
[Music] and one of the efforts to communicate
with Xerox was a set of stuff called Penry papers and I wrote one called
display transducers and this is the drawing that I made for it showing two
ways of doing a display transducer right now and I won't bother going through the
features but you get the idea should look something like that and Xerox had a
projection technology a light valve technology that had some promise back
then so I threw that in also now this is
an important point I think the most misunderstood thing about the Dynabook
is this idea that it's some kind of box it isn't a box and it isn't a piece of
hardware what it is is a service and so the idea back then was it's a Dynabook
if it gives you your information services wherever you are on earth and
so there are lots of different ways we used to say we don't care if there's an
atomic powered computer on the moon beaming down computations as long as you
can compute wherever you want so this is an idea based on Ivan Sutherlands head
mounted display and realizing that that was a pretty good way of doing it in
fact that's something that should be looked into today and then the next effort at Parc to get something built
was this design idea called mini-com and this was done in conjunction actually
with Gordon Bell and Alan Newell so they remember very well we sat in this very
strange panel at a fall joint computer conference talking about various machine
futures and I went through this machine showing how you could package
a thing like a link but a little bit better into something real small and it would be really powerful you do great
stuff with it for kids and herb Grolsch
who is an old warhorse got up and denounced this whole idea as being
absolutely ridiculous and Fred Brooks who was the the fourth person on the panel said to herb he herb said I've
been in this business for 20 years ah and fred brooks fred brooks said herb
i'd like to make an ad hominem remark he
said most of us don't take nearly that long to make up our minds now of course
i knew that this meant that my ideas were good because being denounced by
herb gross has been a sign that your ideas are good i at the same time that
that was going on a group led by Bill
English who had been the main technical guy for Doug Engelbart was building a
system called Pollos and this is the original design of the polos terminal
and that will fit into the story I think Butler and Chuck will have a
good version of it here's the angle Bart Mouse which was great because it existed
but it was very hard to draw with and I think the I think part of the problem
with it was that nobody really wanted to draw over there and it was perfectly
fine for everything else polos group
wanted a different Mouse this is sort of the Chrome and tail fins Mouse but it
had the same basic drawback is that you couldn't draw with it and finally we
built this this was an idea of bill
English's and Roger Bates $200,000 worth of hardware see Butler Lampson Bill
English Roger Bates Roger built the thing this is the thing called the old the old character generator 40 megahertz
of video back in 1972 and you could do almost anything the television screen
could do
driven by ANOVA and you could capture images by this television camera here's
what the that thing looked like back then and here's an experiment to compare
drawing something in a notebook it's a
design of an autogyro that I did and comparing it to the best we could do in
painting on the system and this is the first bitmap painting system done by
Steve Purcell now we get to this this
middle part because this is this is the pivotal point on which I think most of
our ideas revolved and that is that at some point we had to grapple with the
idea that there were going to be real people using this stuff and we had already committed ourselves to children
and when we talked about it among ourselves and in the spring of 72 our
cells were just a few of us Chris Jeffers myself John Jacques we had no
idea what was going on and here's an illustration of why this is difficult
now if you look at this picture for a while you'll see something wrong with it
everybody see it what's wrong with it
teeth are upside down what else whole
mouth what else yeah okay well let's
take a look at it right-side up
now notice that even after you knew what the trick was you still had an emotional
reaction when I turned it right side up that's significant I have to put it back
here to talk because people won't listen to me if I and the the unfortunate thing
seemed to be that in order to understand what's going on we have to admit
something that none of us want to admit and it's not that the universe is not in
our control that is awful but we usually have forced to find that out a lot
earlier but we have to admit something almost a bit as bad and that is that we
don't live in reality but in a dream of our own fashioning a good one for people
who doubted this is something that's good to do every morning along with your
jogging you take your thumbs or your to quarters or an orange you can do it
right now it's okay cameras are on me and you hold one as far as away from me
as you can and you put the other one about halfway there and compare the two
sizes of them don't squint just look at
one and then the other and most of you will see that the one that's further
away is about 80% the size of your thumb if you do it with quarters the quarter
will be even closer in size and if you think about that that can't be because
that image of the smaller one on your retina is exactly 1/2 the size so what's
actually going on is you're not seeing what your eyes are seeing but you're
seeing a reconstruction based on your beliefs of the world and it's that
reconstruction that allows theatrical performances to work and it was
understanding that user interface is basically theater that advanced us quite
a bit of the way now oh I should tell
you that let me tell you the final part of this story is this this illusion is
actually known about physiologically on
the under right-hand side of your brain where you worry about faces there are
two completely distinct little modules one module where ease about face like
thing that's the thing that allows you to see faces and clouds and scares children in
Twilight and so forth when they shadows look like a face and they think it's a
monster that's reporting an upside down face like thing about an inch away a
separate piece of brain tissue is only worried about eyes and mouth and it's
saying well that's an okay mouth and those are okay eyes so there's very
little dissonance here when we go to this guy the face recognition guy is
saying that's an a right-side up face like thing and the eyes and mouth guy is
saying something is very wrong here now
you can imagine why evolution might have arranged for you to worry about this because it might mean you're going to
die in the next 10 seconds
now I can I can tell you it also it also works on mammals also works on all
mammals and if you'd like to try it out on a dog I suggest you pick a small one
okay this is the place that's going to start the third part because that is the
first image ever to go on in Alto and let me now continue with this our
misunderstanding of how the mind works
what I'm presenting here are not my ideas but once I got this strange idea
that the the are mentalities bite me modular it reminded me a lot of reading
Piaget which I never really understood very well that is PJ's fault by the way
in case you don't feel bad when you read Piaget and find it bumbling verbose and
obscure because the fact that's the way he writes read somebody else's summary
of the stuff the ideas are kind of good so Piaget had this idea that there are
three main stages of development in children a sort of a doing stage which
he had a long complicated word for an
image stage in a symbol stage the doing stage is where an object is to grasp it
in a hole is to dig it the image stage is where the kid gets the water pouring
wrong you know you poured from the squat glass under the tall thin one and the
kid says there's more water then Piaget says around 11 or 12 the child starts
dealing with pacts and logic which i think is a special property of Swiss
French children [Laughter]
I've always I've always felt it's around that age that logic fast disappears now
Jerome Bruner now Jerome Bruner is was
the person I think of all of the people that I read that influenced me the most
he wrote a number of books that are classics unlike Piaget Bruner is one of
the classic writers of prose in the English language and around the early
60s he was pushing the following idea that instead of being stages of
development what we've got going in children are actually separate
mentalities and a change of dominance so
one of the experiments that Bruner did is to take a kid who would do the water
pouring thing wrong he would say there's more water in the tall thin glass and
then Bruner would immediately cover up that tall thin glass with a cardboard
and the kid would say oh wait a minute there must be the same amount of water
because where could it go and Brenner would take it away and the kid would look at the glass he said oh but there's
more look at it and Britney would cover it up again the kids would say oh there
there must be the same because where could it go so if you have any 10 year-olds you'd like to tour man
bouncing them back and forth here and you do it partially by blocking the
visual field and so Brenner had convinced himself that there are
actually separate mentalities with separate rules and of course there are people like GaN brick and Arnheim who
thought the same thing about the way we look at images and do visual thinking
that Brenner built a curriculum around it and the curriculum had several
interesting ideas one was it's really great to learn things in this order
because even though the symbolic one is
is the most powerful it's only powerful when it's in the right context in other
words what Bruner said is that logic is a weak method it's really great when
you're in the right territory but it stinks when you're aren't because it
forces you to build up Chains of things that don't lead you anywhere well in
other words if you want to multiply two numbers and 60 ad you're in trouble you
want to find out where more is gonna go in 1325 you're in trouble you have to be an incredible genius and
work with your Roman numerals or epicycles in order to get something
simply because the context is very poor we call those ideas choosing the right
representation in computer science I'd also read a book by Hadamard who did
a survey of the top 100 mathematicians in the world and being French he
included himself in that top 100 but in
fact he was I always thought he did this in order to atone for having invented
the Hadamard transform and forced us all to learn it but his survey indicated
that of these top 100 only a few claim to use symbolization at all they all
claim to use visualizations of some kind an amazing 30 percent including Einstein
were operated down here I in Stein's quote was I have sensations of a
kinesthetic or muscular type the Einstein actually could feel in his arms
in particular the spaces that he was dealing with so this Hannah Hart's book is the cliche
genius as the ability recapped your childhood it will or you can look at it
from Bruner's point of view and say that people we think are creative are able to
think in more than one scheme and the
most creative people of the people who can move from one scheme to the other I
also this also reminded me of a lot of musicians that I'd read about that like
great artists conceived their works down here where the blood and guts flows and
then in our civilization they bring it over here where they where you can apply six centuries of technique to make it
the most beautiful thing that can be imagined so this notion of being able to
move from one to another was very powerful and also this idea that when
you try and explain things to people the weakest way is to tell it better to show
them that's why we started giving demos right away but much better to get them
to do it now was something that we didn't do so well at we didn't sit them
down often enough in front of the terminals themselves and so they could get this visceral feel until many years
too late so this idea that there are
weighs in leads to this little experiment you can try this is inspired
by Seymour Packard and that has to take a child in each of these three stages or
mentalities and get them to try and draw a circle and logo and pampered had done
it with very young children by just getting them to close their eyes and say
make a circle with your body so that kid would start going like this and if he
didn't fall off the stage you could ask the child what are you doing Johnny and
the child would say well I'm going a little and turning a little over and over if you type that into logo you get
a perfect circle that's because the five year old knows differential equations
the circle circle it has constant curvature and so the change in curvature
is zero and Patrick explicitly made logo eco egocentric inertial coordinate
systems in order that so the chip the kid could play the turtle be the turtle
and use what his body his body knew what his mind did not if you will if you do
it with a ten-year-old you get a much different set of results the
ten-year-old really doesn't want to do that ten-year-old tends to be very
visual and so a way to get them to do is you give them a compass and let them
draw lots of circles and after a while they decide gee that the compass is a
constant distance wise are the points of the same distance apart this ought to work and so you pick the pin up and go
out to the rim make a dot you come back turn one degree round you go and you get
a perfect circle locus of points equidistant from a given point now of
course I wouldn't be telling this story if the fifteen-year-old succeeded fifteen-year-old is in the symbolic fact
logic stage and knows the most horrible fact ever discovered by man
[Music] wrong wrong coordinate system and even
if it were a good coordinate system it has very little operational significance
so this led to a slogan I'd like to do
things bye-bye slogans I later discovered people used to complain to me
that it wasn't being professional I finally found that Francis Bacon in the
new organa which has to be a classical enough text to inspire anybody 15 35 or
something Francis Bacon pointed out there are two ways you can explain things one is in prose which in which
you essentially defeat involvement by explaining everything the other is by
form of an aphorism which brahs people in and tries to get them to complete it
themselves and so my characterization of this is the point of view is worth 80 IQ
points right in other words if you're in
the right context you act as though you're incredibly smart and able if
you're in the wrong context you don't de Buono has an image of if you're digging
for gold and you've dug down five feet and you
haven't found it one of the things you can do is dig twice as fast that's
American business when in trouble redouble your efforts but in fact you
might be in trouble because you're digging in the wrong place you might be doing the wrong kinds of things you
might be having the wrong kinds of thoughts so maybe you should shift your context so in anyway this this idea set of ideas
again I want to say that we didn't talk about things exactly this way we read
Bruner a lot and in fact it was Adele who solved how to use Bruner with the
children as you'll see in a minute that led to a slogan which is boy wouldn't it
be great if we could tie these three things together doing with images
generate symbols this is the Bruner direction you should do concrete things
concretely so we used to say simple things should be sent should be simple
and complex things should be possible so it's perlis had this great line once a
programming language that does nothing simply is not worthwhile learning for in
order for doing something complicated so there should be some commitment to
some point of view that gives you leverage my final example is something
that came about a few years later but still in the scope of this and that is
the the following that a guy in Malibu
California wrote a book called the inner game of tennis and in it he purported
that tennis was actually rather easy to learn and the fact what you should do is
not think so hard about it not worry and so forth just go out and do a few things
you should be able to learn how to play tennis in just a couple of hours and that enraged Harry Reasoner who was at
ABC in those days and I think Harry reasoner's from someplace in the Midwest and knew goddamn well at tennis was not
easy even trying to play it for for 25 years and so he sent an ABC News team
out to do Tim Galway in they picked out
35 people and got them to sign legal affidavits saying they'd never played
tennis before and they picked the worst one of these who was a 55 year old lady
40 pounds overweight 5 foot 2 in a muumuu and in front of all of these
people on live television they put a clock at the side of the court and said to Tim Galway you have 20 minutes to
teach her how to play tennis and the next video selection will show you what I mean now
if you're interested in user interface and teaching and stuff like that is I
think you must be if you're here take a look at this user interface
flying through the air I want you to have to watch it the very instant the
ball hits the court I want you to say bounce and the very instant the ball
hits the racket I want you to say hit you just be saying bounce
hip all you need to do is watch the ball okay bounce the key of all the exercises
in the inner game is to focus their minds attention somewhere where it will
not interfere with the body's ability to hit the ball automatically you stand
here and Tim will throw you a couple and you just continue doing the same thing
China hit you not at first at first just say bounce the ball bounces and hit when
you would hit it but don't hit then after you feel like I didn't it go ahead and hit it okay
watch your right arm here bounce yeah something's trying to hit the ball right
when the players first learning the game of tennis and what he says bounce he's
also out of the corner of his eye watching me hit the ball so even though
he's not trying to see how I get it the image of me hitting it is going right
into his memory enjoying a job yeah the
cardia principle in any book written on tennis about a tennis is to watch the
ball no I really sees the ball no the
reason is they get bored so I've tried to give them something to see in the
ball that's interesting so we watch the trajectory of the ball the the
gracefulness of the line made by the balls that flying through the air oh my
get the same put the mind somewhere or work and stay calm and not get you up
tight but relaxed and interested then
the body that's so beautifully don't beautiful come on let's move to the ball boys
listen to how it sounds in your racket hear that yes
[Music] I heard that down like very dull it was
like wood that would sound people this
man Daniel you'll hear the sound that your white Britain
[Applause] be aware that racket is part of your arm
feel uncomfortable not every time I did
start to think things went wrong just stop thinking the body seems to
know what to do you feel with your left
hand where it is how'd it go I'll just win the back there feel it
we're gonna do one other stroke now and
this one's called the server and the server is dis like a dance he just like
a dance we'll come back here I'll show you how to do a dance call the sir okay I've done this dance a
little bit so I'll just do it okay in
fact you can start humming if you want the rhythm okay now we do want you to
stand there shut your eyes shut your
eyes and imagine yourself doing that same dance let's start humming a little
bit as you're doing it yeah
okay now there's no thinking just go do
it quick no thinking just go syrup yeah sir
don't wait on busy keep serving keeps their own thing dududu quick that
doesn't stop fine again don't worry about hitting now notice the wrist snap
oh there you go I get them to hone the
rhythm of my serve so they're not thinking oh this is what are you doing with a job
what are you doing his hands and not trying to remember all up they start
humming and they see themselves doing very similar things muscular movements
to what I did but they're not thinking about it so it just comes out naturally and fluidly that didn't take long
that's called serving oh she didn't
double fault the entire match no need of
it this is a thirty ball rally and became after this
it just started just sort of like floating along you know and doing what came naturally let me know and I said I
pointing toes it's really really beautiful I have a clearer idea of what
you want then let it happen don't try too hard now he sings a couple
Adder I'm gonna try to make you Miss Molly it's going on too long so I'll
just get a little bit harder you just
keep watching the ball
[Music] [Applause]
well when I saw that my reaction was holy shit but in fact that's what we
were trying to do because one of the things you see here is that as Tim
Galway like to say beginning tennis is called chase the ball when you're
chasing the ball you're not doing what you're supposed to be doing when you're
playing tennis so his feeling is you should do anything that you can to catapult a person as
quickly as possible into the lower ranks of the intermediate which is what he did here so that most of the time what she
was doing is doing the kinds of things that real tennis players do and also to
get involved in the actual process of joy because you don't play tennis if you
don't like to hit the ball you can also try and win but mainly it's about
hitting the ball and moving your body and those kinds of things and he gets her involved in it right away
the whole idea behind the kinds of user interfaces that were done at Parc was to
simply trying to get past all that bullshit at the beginning there's no
reason that you have to learn ms-dos to get rid of a file or move things around
or find an application or any of those things those aren't intrinsically hard
there's no entropic reason why you should spend a long time doing those now
what we didn't do so well at Parc either in CSL or SSL is find that way of
getting from the concrete operations to the symbolic it's one thing to delete
three files by dumping them into a trash can but when you have to delete 53 files
things aren't so much fun anymore and it's that bridge of finding ways to
teach the person the symbolic ways of doing things while they're doing the
concrete operations that I think is one of the cruxes and user interface design
today now here we go into the last part
this is a photograph I believe as far as I can tell it it's marked April 72 73
and that is exactly the right time for it for
this to be the actual photograph of the first alto screen I think is the very
day that Steve Purcell got this image up there no not more than a few hours after
Butler and I mean Chuck and Edie and Larry had finished debugging the thing
so that was the start of an era of the
most tremendous fun I've ever had in my life and I just want to show you a little bit of it you get an idea of what
we're all after so now let's start on
that tape B and the challenge was why can't if Peele at Packard can do it why can't
wait because we have a few more functions in mind but still the the HP
35 was there to inspire us I also discovered that one of the things about
having an HP 35 is that it's very stealable now here's this is the
character generated I mentioned built by
Roger Bates and designed by Butler and Bill English and this is Ben laws font
design program one of the best ever done doing an A and Lydian cursive and I put
this up this is the exact same thing as I showed you on a slide so I'm trying to
invoke your imagination as to how wonderful this really looked in person
here's an interesting thing what do you do when you do a rough draft and you're
printing it out in Times Roman and the answer is you use your own handwriting
I never understood why this didn't catch on I think it's a logical consequence of
font mania so again the text can be
rendered very nicely in handwriting here's a phonetic font that we thought
teachers would might be interested in designing from Pittman initial teaching
alphabet I never thought much of it but
one of the nice things was that we didn't have to tell teachers what to do
on this system now here's the world's first painting program done by I
designed it in the spring of 72 and
Steve Purcell that most amazing person
did this while he was us I think he's a sophomore at stay
and at this time I'm not sure you'd ever written a program before in his life
I never asked anybody that I hired whether they'd programs or not so I usually didn't want to know and the
other thing you notice is that the control here is being done by a
character recognizer and that was done by none other than John Jacques that
proves that venture capitalists can learn how to program so most of the
things that Mac Mac Paint could do can do this this system did so we see the
different brushes and you can make up your custom brush shapes right there
you'll occasionally see the symptoms of an analog mouse here I believe
occasionally will jump a little bit make
small brushes and so forth I should
mention that in these early things you'll see some flicker because I shot all of this early stuff on a Bolex and
16 millimeter so we didn't have any video equipment back then specular
reflections and let's see I think we make a brush here and do spray painting
no I guess not there we are
so this is well running by the summer of 72 I should mention that the one of the
reasons why most of the stuff we did at Parc worked out there's the Cookie
Monster that I drew this is a Xerox executive he was our mascot Winnie the
Pooh space war practically drove small
talk into existence we wanted a space war to be half a page and here's some
early animation from I think around August done on the old character
generator and from all of this simulation we got in an idea of just how
much power the Alto would need to have to do all of this stuff for real I think
the next segment is an example of some music I did the music programming on the
Nova Nova 800 was a pretty good machine
and it could do the following this was keyed in by Steve Saunders
now of course what we didn't want was just for voices like that Brandenburg we
wanted 12 because 12 or what you need to do things like organs and other keyboard
instruments lower in stature than the organ and so we sort of multiplied what
that could do by by a factor of three or four and multiplied the animation by
that and we'd also done some studies on what the dot size of the display should
be we discovered that the that there was
a completely nonlinear transformation in the eye that there was a place when you
went from fifty dots per inch to one-hundred dots per inch which cost you
four times as many dots which was pretty expensive but it was ten times harder to see that they were dots if you think of
it the last thing you want to see on a display is a dot if you think of that
there might be a maximum in that curve it happened in corn sweets book when you
work it out for video displays that the worst possible display resolution that
you can ever choose happens to be fifty dots per inch now you might wonder what
IBM's resolution is it happens to be 50
dots per inch now let's continue I want to now mention that there are two people
who really you know I'm this is the last
of my stuff here from now on the group
was run by this wonderful person
notice the pigtails folks this is and
this is the first kind of stuff that kids did in small talk small talk was
done by the second great person that's Dan Ingalls without those two people nothing would
have happened I did have one little thing to do okay stop it for one second
please I didn't have one little thing to do with small talk I wrote the original
evaluator but I neglected to leave out
things like the storage allocator and things like that and it came about as a
result of a bet I said well you know you can do an arbitrarily powerful
language in just a half a page of code and unfortunately they called my bluff
and I had to figure out how to do it I finally wrote that half page of code and
Dan went out and implemented it for real and it took about seven hundred lines of
basic but I only wrote a half page of code I want you to know that and when
the stark weathers laser printers started working people used to accuse me of writing my page long programs in
pocket Bible fonts now this this next
sequence is I don't know why we actually shot this but this happens to be the
first time a window has ever done an overlapping window has ever done on an
alto I wrote my usual one-page of small talk and Diana Mary sweated for months
doing all the code that I never wanted to see in order to do the following okay
let it go now in those days I felt
strongly that windows should be draggable I still do damnit so the idea
was if we could ever do animation reasonably we should be able to do this
in real time so notice that that text is overlapping there this is like this is
practically like the first day it worked and now we're going to display you
notice the other guy is popping up in front this is one of those little ideas
that people have in the shower here's
now we needed some frames and so this is a little frame program using the great
ideas of having little moving handles and stretching handles to go around this
stuff I once asked you know as I always
found it strange that Park was willing to put a twenty thousand dollar machine in my office but wasn't willing to put a
thousand dollar shower I explained to them I got much more done in the shower
than any other appliance nobody would ever believe me
showers now here's here's a great
sequence this is Larry Tesler Smitty mouse editor which he did to convince
the polos people that modeless editing was the way to go this was written in
small talk and this is an example of very early programming in small talk and
it's to do the list a cons class so what
you see there is a message called first time which when it creates the instance
is going to look for a head and a tail that's like cons here is if you see is
followed by question mark then return
the answer list if you see is followed by list then return the answer true so
what we're typing in here are responses to messages that any member of this cons
pair might have and if you see anything
else return false and if you see the
message head followed by in ass and arrow then rebind head so that's the
same as rip Lakha and if you don't see the arrow then just give back the answer
head so that's the same as car and here if you see the message tail followed an
arrow then give back tail that's rip lacta and if you don't see it then give
back tail that's cooter and that's all
you have to do to create an infinite number of Lisp cons pairs I think you
can barely see some on the side when they put this thing to tape they
unfortunately cut off some of the stuff so you notice this editor you're just
moving around you'll notice him be able to go back to this transcript so he made
himself a list there and called X and
now you can't that's too bad it's off the screen I have to redo that but what
it's giving back is B and now he's putting a C in there
and now he decides well I don't really like to see those that stuff printed so
he goes up here and he says look at the small talk evaluator and what gets
printed out is this loop repeat read
something give it the message of Al and then give it the message print and what that means is that for any new class we
make all we have to do is give it the message a response to the message print
and it will print automatically so that's pretty neat print is thus generic
print is distributed over all the classes in the system so here is this
guy's local thing you notice in Larry's editor you just go up and fix the thing
that you're working on put a few more things in there and that's not quite
right so you go back and change it so you want to print a left parenthesis you want to
send the message print to whatever is in the head you want to print a period you
want to send the message print to whatever is in the tail and print a
right parenthesis then if you go back down here and Larry's editor you see it
prints out a dot B which is what it's supposed to do then he says oh well why
not go in here and give it a response to
the message length and that is if the
tail is a list then we want to the response is 1 plus whatever the length
of the tail is if the tail isn't a list
then we are the end so the length is 1 that's all you have to write there so if
you type that in back comes the answer and if you go down and make a little
more complex to cons C to the existing
thing X you get C dot a dot B and that you can't see it there but length is
giving back 3 so that's a quick pass at small talk and that is the language that
the children programmed in everything was done by writing those class definitions now the meanwhile per cell
was not idle
now Chuck in Butler both complain that the alto is slow and that that is the
only dispute I have with him that is absolute bullshit look at this this this
machine could do 120 square inches of graphics in arbitrary freeze it for a
second please 120 square inches of graphics was about that much graphics you can divide it up
anyway 80 ping-pong balls or were in number of Pegasus or something like that
in two and a half D with transparencies
at ten frames a second that's what the alto could do and it took 300 words 300
instructions of micro code to do that and I believe that Purcell had the will
part of his brain to Faust but aside
from that really worked well okay so
that was an example of the alto really moving along and in fact I think just as
a comment here I think that the if we had only done the following on the alto
which is to added 1k of microcode RAM a year we would have been in fantastic
shape because that's what we ran out of let's go let's go further I think the
next sequence shows that music was coming along
here's a piece written by Chris Jeffers and round this time we moved into our
playpen downstairs across the street at
what used to be the singer building this
piece is called a happy hacker [Music] there's the organ that's playing the
piece [Music]
this is great stuff [Music]
now here comes Adelle and Joe now if we
always believe that Joe was some secret lover that Adele had had but in fact we
discovered that it was a little box that
she gave give a name to and she came up with a way of teaching this very abstract high-level language that the
Japanese are very interested in these days two young children by giving them an
already written class called box and you could make a box and give it a name and
turn it and here's Jill and you could
turn her and you could write and right
from the beginning the kids would start off doing things that involve more than one thing more than one instance of a
class so here's one where we turned Jo
one direction and we turn Jill in the other this method of teaching the
children was so effective that after the first couple of bunches Adelle was just
able to leave her little box book around and the children would pick it up by
themselves because basically all you have to understand about small talk is
that you can make instances the instances can receive messages and you can add more messages to the class
definitions so here's a more complicated repeat loop that grows one and shrinks
the other while it's rotating them and
this is when things really started cooking because we were able to instead
of doing turtle graphics which never leads to a tool we're able to get to
what one of our major aims was was to have children be able to do their own
applications programs Adelle organized all of this stuff I was sort of in the
background helping a little bit but she conceived of all of these programs she
wrote the curriculum she just did everything here I think is the first
tool ever done by a child this is Marian gold Dean's painting
program she had seen per cells painting program so she decided to write one herself so there's a menu up there of
shapes that you can grab those smaller ones are brushes there so
she now you can imagine behind this that
there's actually a single class called polygons that every time she points into
one of those is making instances that can splat up on the screen the polygons
can be called with making with four to make four sides and it can be called
with six to make a hexagon and so forth so there's very little code here and in
fact this program is not much larger than it would be in logo but no logo
child has ever done that program because it's not at all obvious in logo a part
of the thing with novice programmers is that they have to use their short-term memories for all kinds of things and you
have to have some way of allowing them to see how structures go together here's
some very young children that we did in collaboration with Radia Perlman from MIT and they needed big buttons to push
and they could drive the turtle around the turtle could remember things and so even very young children could write
procedures and their big problem was doing planning the music editing system
is under constant revision in an attempt to find a comfortable user interface for
editing notes and toddlers this is an excerpt from a movie we made about
computer music the star Rachel has played piano since she was five years
old [Music]
so it's being captured [Music]
she corrects a mistake with no it's okay because you can edit that transcript at
age six rachel begins where's her
keyboard technique their philosophy is to introduce children early to music and
art skills developing the necessary motor skill dexterity and hand-eye
coordination and making the learning of these skills as natural a part of the
child's development as our language skills so she's played she's played and
design skills shut up Dianna she's she's she's played
in it in a compliment now she's gonna play along with the accompaniment now
this is sort of our idea what a convivial tool really means
[Music] now of course Rachel was just an an
average kid this is a nice thing that
Adele organizes Marion was had an
enormous amount of hutzpah and she decided it would be great to teach a
class of students so we videotaped literally every second of this
and we learned a lot from this now here is Susan Hammett with her escalation of
the painting program notice there's a menu down here now and this is sort of
her version of what we now think of as Mac draw of course Mac draw is 70k of
machine code and you have to buy it but aside from that this is very similar so
notice there are little birdies that came up there and she's changing the
color and she's going to change the size
of the thing and now she's gonna move it
she's gonna make another instance of it
she wants to select one of them so she puts out the birdies or handles there
now she's going to make it very small because if you make it small and make it
with a lot of sides you get a circle and
again she could make instances of that well she changed the color now she's
going to make instances of that and so
this is an applications program done by a 12 year 12 year old child and it is
not large let me claim the generic one
page for it it must be in some font
actually it's quite a small program so it's really only one class definition
and a menu she did the menu menus weren't built in now here's Steve putz
who was unfortunately grown old like the rest of us he's actually in his 20s now
at age 15 he was interested in ham radio
and he hated to design to draw a circuit diagram so he did this system which is
pretty damn sophisticated it has a menu
like they're a pop-up menu for often
used things like open and closed dots and has they wants that right where the
cursor is right where he's looking and down below he can pick up components and
as you can see rotate them as one of my favorite programs by a
child because less than ten years before it a PhD was awarded for a system that
wasn't as good so he felt at this point that there is some hope for progress in
novice programming
and the final feature he demonstrates here in a second is being able to float
text wherever he wants it in the drawing
this is about all the stuff I'm showing you is done between the summer of 74 and
the end of 75 and finally he types in
the text
I was just around this point that we started seeing where small talk started
running out of steam and part of it was a design it was a profit here's Bruce
horn I think this is the first time this
is done by anybody in real time he's 15
years old also here yeah that's off the
frame there but you get the idea it's pretty damn good
now here's the the tomber editor there
were a lot of people involved with music Kris Jeffers myself
Ted Caylor Chuck Thacker helped a lot in
the early stages of it Bob Shore Steve Saunders who figured out how to do
Channing's FM with without using a multiply which we didn't have the alto
could do 12 voices in real time of the kind of stuff that the curse while does
but do it better it could do eight voices of real-time FM and I could do it
without any hardware except additional hardware except a D to a converter this
is another 300 words of micro code so you take 300 words here and 300 words
for the animation system and add another thousand or two for small talk
you can do pretty unbelievable things on an alto I think it's still my favorite
favorite machine I've ever ever used it was just sort of right so here he's
changing the pitch between those two
bars to get a vibrato and now he's going to a different Tom Braille palettes if
you will here's woodwinds but what he really wants is something
like a bowed string like a cello which is something that FM can do really well
then you can play a piece that you've played in or edited in in real time each
voice with a different timbre this is
one of our favorites is the fugue from the passacaglia and Fugue in C minor
so the idea was on a personal computer you should be able to get all of this
function without there having to be any hardware specially prepared except
things that you always have to have as Wes Clark points out like D to a and A
to D converters for God's sakes but all the rest of the stuff should be done by
software remember a hardware is just software crystallized early and what we
really want is to crystallize our functions late now here is what was made
out of per cells animation primitives system calls Shazam then by some
visitors Ron Becker and Eric Martin Tom Horsley with some assistance from
Purcell and some of the gang doing what every animator always wanted to do which
is to draw on a cell there on the right then the left the movie window just pick
up that thing you've drawn and shown it what showed it what to do this is a
wonderful system because with per cells microcode primitives the whole user
interface for this was only five pages of small talk they did four different
systems that summer of 74 which is when this was done so notice he's single
stepping it because what he really wants to do is to get a squash effect down
here so he's going to zap that guy there
and what he's going to do is to is to tie that space in the frame into another
animation cell which is transparently superimposed over the thing on the right
there so when you see that menu come coming up it's actually for the the
transparent cell and what he's going to paint here is not going to disturb what's there he's just using it as a
reference and notice it's being inserted in in real time in fact you could do
dynamic graphics inserted in real time what the animator wants to see is how
many features of the image does he have to include before you get the illusion
of single frame squashing so as he puts
in the specular reflection you see that does it just one frame changed but it
looks like it's squashing and just to
give Steve Purcell a few more plot it's one of the problem with doing a system like this is that
white is ambiguous that can either mean
transfer here in the thing like a window frame here can either mean transparent
or it can mean opaque and the default
was transparent so here he's got a a
blob animating there and now he's gonna bring up this guy and just swing it
around so we can see that the holes in it are actually letting us look through
so now what he's done is gotten this is
the painting window and now he's going to opaque paint and he's gonna paint
over those notice the painting is being simultaneously looked into the other
window there and so now some of the what
was white transparence is now white opaque and now he's going to pick up the thing and moving around this was all
done on the alto amazing one of the one
of the great stories of course is the great leaps backward possible in
research in fact we never did anything like this as good again another one of
our famous twelve-year-old girls who loved horses as they often do Shazaam
was bigger than a twelve-year-old can handle but we discovered that kids can
read much larger programs written by other people and make modifications and
so she was able to add a feature to this animation system that she wouldn't have
been able to program herself
I can superimpose the jockey this is the feature she added which is to allow
images to be combined into single animated images just like sketchpad and
now she has a composite
okay let's I think there's there's a pause in there well this at the end of
my talk except for one announcement one more little show and that is I didn't
know how to talk about all the stuff we did at Park there are many years of work
afterwards and in fact I didn't know how to really talk about the Dinah book of
the future I still think that as with Bush's original vision if somebody just
sat down and implemented what Bush had wanted in 1945 and didn't try and add
any extra features we would like it today I think the same thing is true
about what we wanted for the Dynabook if we only had what we wanted then in one
machine we would like it today now of course that doesn't mean that the Dinah
book covered everything one of the things that I didn't anticipate was what
networks really do to you and that is they change the whole environment that
you're in they don't just hook environments together and I subscribed like everybody
else because we were all old time sharing hackers to the idea that a
personal computer is always connected but in fact the very connection of it
through the ethernet changed everything the ethernet dominated the press
protocol for printing dominated all the machines that were on it it was the
thing that was eternal and the services that provided not the machines that we
built and the kinds of services that you can get through networking are so large
and so vast that almost none of the browsing techniques that we originated
at Parc are going to be of any use in five years so there's much more to come
in personal computing than just the Dynabook but I'm gonna stop here with
just one look ahead and it's a look that
we didn't do it park but one I got from a person who worked at Parc Alvy Ray
Smith wonderful person and a not yet another graduate student of Dave Evans
IDI Catmull and this is a special effects sequence done at Lucasfilm and
it returns to my original theme of Fantasia and everybody should be able to
do their own and the Fantasia here is that there is a gadget called the
Genesis bomb the Genesis bomb can be exploded on an airless planet and a fire
races across the surface of the planet and converts the planet into life
and of course the the the dichotomy of
the thing is of course if you explode it on a planet that already has life it
converts that into new life too so it's both the ultimate life giver and the
ultimate life destroyer they wanted to show what this would look like they went
to Industrial Light & Magic and they said oh now that's too complicated for model building too much detail we can't
show mountains rising and waters coming up and they went to the animators the
animator said oh no we can't do that there's too hard to do it cost too much
we can't get it done in the Philippines and so forth and so the computer science
group at Lucasfilm said we'll let us take a whack at it and here's what they
came up with what I have done is I hated
the original music in the film so I just deep-six it I found so much better stuff
written by vogner and also I mean if
you're gonna steal you might as well steal from the best so I say never steal
anything small and make sure the lie is large the other the music though is also
synthesized by computers so what we have here in a minute and 10 seconds is a
vision of something that if we were to increase current day personal computers
by about the same ratio that the Alto increased cycles from time sharing we
could do in five or six years so this is
definitely on the horizon and this is I claim is what we all at Parc had in mind
and I don't think you can argue with me but you can't prove I'm wrong remember
the truth is but a lie that hasn't been found out yet okay so let's roll this sequence
[Music] [Applause]
well the Greeks held that the visual arts were the imitation of life but the
computer arts are the imitation of creation itself thank you very much