| 
  • If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • Stop wasting time looking for files and revisions. Connect your Gmail, DriveDropbox, and Slack accounts and in less than 2 minutes, Dokkio will automatically organize all your file attachments. Learn more and claim your free account.

View
 

FrontPage

Page history last edited by Bram Pitoyo 8 years ago

Verbatim (& highly snarky opinionated) notations taken at 

Engelbart and The Dawn of Interactive Computing

held at the Stanford University

on December 9, 2008

1–5 pm

ON

“The better we get,

the better we get at

getting better.”

 

 

Robert Sprouli, Vice President, Sun Microsystems

 

Allen Kay famously said that "the best way to predict the future is to invent it." Computer mouse inventor Doug Englebart and his team predicted the future by demoing it. Their work not only resulted in the mouse, but a style of interacting with data that created a new way of doing work. Doug may complain that some of the most important vision may get lost in translation, even dumbed down for us ‘ordinary mortals.’

 

I would like you to think of what the world was 40 years ago. But even more important is to remember what computing was. Most computers was mainframe. There were only few personal computers. I had a fortune to use one. It filled a room, costed half a million dollar, could only be used an hour every day. Not exactly a personal computer. Back then, interactions was slow and cranky.

 

 

Charles House. Long career in Engineering at HP. Executive Director. Media X at Stanford University.

 

April ’64 is the signal was when Doug started the work in this lab. When IBM launches the 360. It shipped a year later. They shipped 118 machine in the first decade of the company. The seven dwarves shipped roughly 2 units every month that decade. I started because of his input because of the HP 1800. The significance to the box for me: was that the forecast was 31 units. ASR-33 was good enough for many people.

 

I had the fortune of introducing Doug Engelbart. I was required to learn before becoming his lab assistant.

 

In March, WANG introduced the WANG 4000 computer systems. By December, they only had 7 orders. Yet their annual reports said “the product has broad potential market.” Clearly, we talked about how computing was not like it is today.

 

Doug is, though, much remembered for the mouse. His purpose was not to invent gadgets, but to have people come together and solve problems. Collective intelligence. ABC network. 93 networks. I’m mindful that we don’t know yet how the collective IQ will work, and it’s also hard to know when exactly it will happen. At ’97, in an ACM talk, there was a panel about the future of computing. The 5th guy was Vint Cerf, and all he wanted to talk about is the internet. Nobody cared back then, of course.

 

 

Q&A With Panel Composed of Doug’s Collaborators

 

Doug was a cool demoer. Wonder if that’s true for the rest of the group. Jeff?

Jeff: everybody was nervous. Computer was actually more unreliable than it was today.

 

Did it crash?

Jeff: yes. We made a program to restore it. If you look at the original film, there is one frame where the crash happened.

 

The demo was 50 miles away. Your thoughts?

Bill: it wasn’t very hard. We had two links.

 

If that had failed, then what?

Then Doug would had a blast in the auditorium [laughs.]

 

What sort of ideas were missing? ideas left out? Anyone?

Jeff: technically, almost everything is in the demo. But the deeper ideas that Doug had, abut bootstrapping, co-evolution of ideas and building tools that build tools, was not discussed. It was a demo only in the technical sense.

It strikes me that our search technologies today are very highlighted. Yet Doug’s system was more about structure than about search. Can you comment?

 

Jeff: I knew for sure that search was in a file. All your info was organized in a chunk, and you can easily set up a regular expression and match.

If you measure the power that is in the iPhone, our system was a mili-iPhone. 1/1000th of the power, divided between 10 keyboards.

The address space was a killer. Having 96 Mb at your disposal is very mindblowing.

 

Andy, I had the impression that these guys didn’t have any idea, and you kind of figured it all out.

Andy: Let me give you my reaction: OMG. I couldn’t bring myself to believe that the demo was for real. So I question him very hard after the presentation. He tolerated it, and even invited me to inspect the system for two days. I was immersed. At the time, I had been working on our first hypertext system with Ted Nelson. So to come see Doug’s system that has unbelievable richness and complexity was mind blowing. This demo, unlike just about anything else, stood the test of time. My takeaway from it was the deep and systematic design methodologies that these people had. I never seen software engineering like that. We worked with the equivalent of hammer and and chisel, and these guys were bootstrapping: building tools that build tools.

 

Jeff, you have a segment in the demo where you demonstrated bootstrapping. It seems that coordinating bootstrapping is tricky because everything depends on everything else.

Jeff: there was an engineer who came up an idea with a compiler who we can build a bigger compiler from, and that compiler will built a bigger compiler. We had a language called machine oriented language 940. Every word was tuned to the machine. Technically, there was also the entire command language specified in this bootstrapping system.

 

Don: We knew that things change very quickly. Rapid prototyping. The compiler trick was one of the tools in our kit bag. Doug talked about the command or control language. Which means that every thing that the user write, the computer does and then give feedback. So we deveoped a compiler to process those codes. But at the same time, the user interface across the system was pretty consistent.

One of the interesting example is how we moved to new computer architectures. When we moved to a new machine, we had to rebuilt from scratch.

This happened with the ARPAnet, when one of the websites was SRI, and so we were bootstrapping the system. We were using the internet to actually get things done, rather than develop for it. The fastest way to get information across the country back then was actually to take it on a tape, and to fly it out on a plane. With this, you can write a major revision and then have it done the next day. This was revolutionary.

The thing with bootstrapping is that you’re actually using what you’re building everyday. Each statement even had an identifier of the author who wrote it, so you literally know who to ask.

 

I get the impression that the demo was the first public unveiling of what you guys were working on. But was there a pre and post-demo at the lab? And did anything change?

I think the demo was a punctuation mark. We had a large projector that you can stop one frame at a time. I remember having to carry that thing across the country and show the movie from it. There is a sequence in there where the cursor is down in the corner, then it moved right next to where the text is in the next 1/24th second. This was how fast the system is. We often got feedbacks like “what did you just do back there?” by people looking over our shoulders. It was so fast.

 

You mention the OMG moment. But was was the biggest learning you took away?

It was a vastly expanded vision. We were already working on the second system, trashing the first specs. So, I think the demo opened up the vision of what could be done while we manage the incredible depth of bootstrapping. We collaboratively did for four years what we called “Language for System Developers.”

This whole project was developed before “open source” was coined. This project was open system. If you look at the system, the way that any editing, searching and filtering was done was through Snowball, which is a regular expression compiler.

 

What about the hardware?

It was prety much the stock hardware. There only some things that was very unique.

 

What about the mouse?

The year was 1963. We had a thing at NASA where we evaluate various pointing interfaces. Doug had a sketch of the mouse. We took that system, and one of the guys said that he would create a box for it.

I went back and read Doug’s 1962 paper about augmenting human intellect. There were 2 parts to it. The second part begs all kinds of question. What was it like to have Doug trying to visualize abstract ideas from guys who were trying to engineer the system and make it work? It’s characteristic of visionaries to have much difficulties communicating and getting the thing to be used by the rest of the work. It must’ve been less than a smooth session. SDS-940 was full. I remember when we were down to 4-memory spaces, and Doug wanted more.

 

Was the system evolved in a slow way, or was there major phases in the evolution?

Before we had the SDS-940, we had the CDC-3100, and we had built some major pieces of the system for that. This was single user. Moving to the 940 was a major breakthrough because it allows for multiple users. Then there was the freezing the code before the presentation. There was frustration over all of that.

 

I’m sure that Doug must’ve been frustrated in communicating all these ideas?

Partly because the machine was so small that we couldn’t put a lot in it. It had to be extremely general-purpose, so it had a very nerdy, general expressions to it. So the answer to the question to “why can’t it do this or that” was because it was made for general purposes.

 

So, then, did the system get bigger or faster?

Yes, gradually it got more reliable, and I think a few more of Doug’s idea got implemented.

 

We saw use of collaborative video and audio. Was that a system feature?

That was pretty much put together for the demo. I had a video mixer. The thing that was done in the software was move the layout to the side, so the picture had a place to show.

 

Andy: I think what makes this demo such a wonderful occasion was that his vision hasn’t really been realized fully in today’s environment. We looked back with a sense of nostalgia, but you ask yourself “what have we got today?” and I would say, cynically, we only have a collection of tools. We have PowerPoint, Illustrator and Photoshop, so we can do the individual tasks better with more functionality—but they don’t play nice with each other. Sure, there is the import/export feature, but only in the lowest common denominator: the bitmap. Back then, it was about authoring, communicating and collaborating.

 

But today we have better search and universal access. This is great, in comparison to older systems. We lost some. We gained some.

 

What I look forward is the integration, the totality of the broad vision. And I don’t see, frankly, how we’re going to get there, because software today are built by teams, and it often reflects that organization.

 

You ask, what about web 2.0 and mashups? I think that the architecture that was defined by these wizards, being in the spirit of tinkering and mashing things together, was pioneering.

 

So, in sense, we can look back to the future.

 

At the time, one of the things that Doug got credit for was persisting when 97% of the public actually didn’t believe. We talked about the problem that we had in understanding Doug’s vision. Imagine everyone else’s!

 

Andy: Let me tell you an anecdote about that: I was invited to a meeting. Doug had a projector. They had the demo for an hour an a half. I thought it was just stunning. And there was this pause after it ends.

 

And someone said: “Doug, I don’t get it. Everything that you could do here I could also do in my teletype system.”

 

And I thought to myself: “if the best and the brightest don’t get this, then what hope do we have?” Look at the multimedia conferencing collaborative that we had back then. We can’t do that today!

 

Multitask. What was it like to work with Doug in a massive integrated environment? 

Jeff: one of the ideas Doug has is co-evolution. We had to invent everything, but as we put things into work, it connects to each other. One of Doug’s idea was bootstrapping experiment. Not just invent, throw and see what comes up, but also see if we can use these changes to invent better and better things. Outside of this demo, I haven’t seen it done.

 

Part of the answer of the question was: did Doug pushed the co-evolution aspect? We wished we did more.

In the last 40 years, we have built computer systems where the structure was the least malleable part. It’s like we wrote the wrong tools. Doug envisioned a structure and parts that are both malleable. 

 

I think that, perhaps, the difficulty is that people are not that comfortable with structure. Some know it right away, and others want to copy, paste, and repair the structure later. When you deal with text and document, perhaps you need to use both approaches.

 

Andy: exactly that. The way that I think was driven too deeply was the hierarchical structure that governs everything else. Some people never used it. Some used it fanatically. The problem is, we don’t have very many choices with today’s system.

 

How far back we need to go to the future. And I think we need to go even way back to the notion of the copyright. Back then, there was so much more notion of the open source. People were more used to working together, rather than figuring out “who’s going to get this part?” partly because there was not a big market for it.

 

Anybody in academia was completely open. And recall that, back then, public money means public ownership. In fact, sharing and communicating information was the part that wasn’t easy.

 

Just to give you an idea of what the time was like. We lived and died because of the circulation of the virtual, funding monies. One of these people was Allen Kay. He went form that to XEROX PARC. A lot of people in the audience was profoundly impacted.

 

The way in which you worked together, and the way in which the tools impacted how people worked together. Did the tools improve your teamwork, and what was the relationship between the tools and the people?

I thought more of hypertext as something that you would use in your own work. So in the following years, we did experiment with my students, and I believe we had the first online communities. At the time, this just didn’t exist. The idea of putting everybody online at the same time with the same materials, commenting on everybody’s idea, was new. And it worked like a charm.

 

PARC had a similar approach. There is a satisfaction of developing tools for “people down the hall.” There was a sense of self-reliance. And this wiss harder to do now: the notion of the team developing the tools for the team.

 

Speaking of a collective intelligence and using computer to bootstrap. I would like to pose a question. How can we put the controls of the voting system from the few to the public. Open source?

I would rule that as a problem of our time. You may have noticed that John Markoff written about how “the thieves have taken over the internet,” when one could make a huge concerted effort to change everyone. We no longer have the luxury of starting small and growing up, because now if you want to change something, you have to change at that scale. The thing that you talked about is very hard to apply for small communities.

 

What did Doug said that you still don’t understand?

Well, there is one thing that Doug had a vision of: the idea that when your media is in digital form, you can really break out of the way that you use to represent things. Text reading in left to right, top to bottom direction is what we’ve been doing for couple thousands of years. So why have the computers do it that way? Maybe there are other ways to do it.

 

There is, in a sense, nothing natural about natural language.

 

We had a system of identifier built in, so everybody has ownership, and the concept was to expand this system to ARPA system.

 

What would a 2009 version of the demo look like?

As you’ve heard from many people, there is not an integrated, knocks-off-your-socks tool. There is some additional works to be done in this area.

 

 

Robert W. Taylor, former program director, ARPA

&

John Markoff, Senior Writer, NYTimes

 

John: Bob was the first funder for NASA for what became the mouse. He then conceived of and funded the ARPAnet. He then funded the demo, and then later, he funded PARC. That’s an unprecendented strength of getting it right so many times. He wrote in 1967 about computer as a communication device. He had a wonderful system of what we don’t even have yet. Maybe we can start by talking about computing in the 60’s. My sense is that you were quite alone.

 

When you came to NASA as a program manager, what were you trying to do?

Bob: In 1960, I read a paper. The name of the paper was human–computer symbiosis, and it opened my eyes about what the computer can do or be. I had a thesis, and went to the computer center for help, and they told me about punch cards. I can’t believe this. This just seemed insane to me. So, a couple of years or so later after reading this paper, he outlined what a partnership between a properly designed computer and human being could be. A couple of years later, a guy wrote a proposal about Augmenting The Human Intellect. So I got in touch with this guy, met in DC, and got the NASA contract. We got him and his group off the ground.

 

Maybe a year or two later, he came to ARPA and started the computer research program. I told him about Engelbart, and he funded Engelbart. Then I went to ARPA, and increased the funding again.

 

One amazing thing about this was that computing on those days was believed to be largely arithmetic. Data, payrolls, numbers, etc. But I wasn’t interested in number, and neither was Doug.

 

Furthermore, Doug and his group were able to take off-the-shelf computer hardware and transform what you can do with it into a software. Software was generally much harder to understand for people, being intangible objects.

 

I’m reminded about the story of a contractor who evaluated the weight requirement of the space around a system. He was going around and “we need to know how much did the computer weight.” The hardware was fair-game.

 

But then he said “well, how about the software? How much did that weigh?” We said “it didn’t weight anything.”

 

He came back a couple of days later, got very angry and brought a box containing cards with holes and told me “this is the software!” and we said “no, the hole is the software.”

 

But it’s all relative, because even the computer establishments in those days was dedicated to batch processing. Computer conferences, for example, were always focused on panel discussions attacking timesharing and the idea of interactive computing. The reasons were many: it was too expensive, computer time is worth more than human time, etc. So the barrier was not only ignorance and opposition, because they’re also opposed to the whole idea of interactive computer.

 

In 1967, I invited IBM to join this ARPAnet experiment. And they said “we’re not interested in computers.”

 

I went to AT&T Bell Labs and said “I like you to be known through this system, and not charge you for it.” And they looked at t and said “packet switching won’t work. We’re not interested.”

 

So it’s very nice that 40 years later, Doug and all that work was properly rewarded.

 

At SRI, early on, Doug was having a problem. Someone wanted to see me, and then said “why are we funding it?” and I said “he was doing something really important and I believe in it.” I got the sense that he felt that he would dissapear along with his funding source. So I did the best I could to justify finding Doug.

 

Even this this community, people who worked in interactive computing, there was a pecking order of some kind. And Doug’s group was at the bottom of that order. Doug and I talked about the demo in early ’68, and he said “it’s going to cost a fortune!” and I said “don’t worry about it.”

 

A year or two after this demo, Doug told me a story: There was a government contracting officer who came around and check on the expenditures. He came to him before the demo, having seen the budget numbers, and said “what is this all about.” Doug explained, and then he replied “if this thing paid off, I’m going to deny knowing all this expenditure!”

 

Doug had an aspect of both collective intelligence and interface. I’m asking because you seem to understand the concept of community early on.

I think that what attracted me was because he had the idea that the computer would augment human intellect. He had interest in shifting and organizing knowledge. Collaborative editing. This thing was more interesting to me than the nature of the interface itself. The interface was no more than a television set. And that television set you saw was actually the shot of a camera looking at a tube.

 

The one story that really struck me was how the obviousness of the idea of the ARPAnet came through. It speaks to where innovation comes from.

Prior to me leaving NASA and joining ARPA. Lick had started a research funding at UC Berkeley, University of California and MIT earlier, and all these places had experimental timeshare terminals built by ARPA, unconnected to each other. This is silly. So this was one motivation why.

 

But I also visited places that we were funding, and we often ran into a set of common interests. So I thought “why don’t we connect these people together?” So we did something we haven’t done before: initiated research rather than soliciting proposals.

 

Every year, we’ll have an annual contractor meeting. The agendas for these meeting was work in the morning, play in the afternoon and work at night. Now, at one of these metings in 1967, I announce for the first time to the group of contrators that we were going to start this research program. When I invited them to become ‘nodes’ in this ARPAnet system, they all had their own individual computer systems, and I offered to connect them together. Well, most of them was skeptical, because they don’t want others to “use of their machine cycles.”

 

Doug, almost standing alone, said that this was a great idea, and even said that SRI will handle the gateway, so that people who want to get information can get through us.

Comments (0)

You don't have permission to comment on this page.