From: Subject: Interview with Ken Olsen Date: Wed, 22 Jun 2005 16:35:28 -0700 MIME-Version: 1.0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Location: http://www.americanhistory.si.edu/collections/comphist/olsen.html X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2180 Interview with Ken Olsen

Ken Olsen Interview

Interview Conducted at
Digital Equipment Corporation

Digital=20 Historical Collection Exhibit

Transcript of a Oral History Interview with
Ken Olsen =
Digital=20 Equipment Corporation

Interviewer: David Allison
Division = of=20 Information Technology & Society
National Museum of American = History,=20 Smithsonian Institution
September 28, 29, 1988=20

TABLE OF CONTENTS

David Allison (DKA): I'm interested in how you first became = interested in=20 electronics. I read that you set up your own radio station. Whether = that's true=20 or not, how did you first get interested in electronics?=20

  • Learning About Radio and electronics

    Ken Olsen (KO): When I was a teenager in the late 30's and early = 40's,=20 electronics wasn't a word. You were interested in radio if you were = interested=20 in electronics. Most of the information came from POPULAR MECHANICS and = POPULAR=20 SCIENCE magazine. The books in the library had very little on radio. I = remember=20 very clearly they stopped at a, the spark _____ transmitter which was=20 approximately World War I time. Anything I could get hold of I seized = and=20 studied and read. And any chance to experiment. If anybody threw out a = radio,=20 which they didn't do very often during the Depression, I, of course, = stripped=20 the parts and used it to do experiments. The experiments were limited = because we=20 could never get enough parts to do anything very exciting. The war = helped.=20

    DKA: Ken, was this something that you did alone? Did you do it with = other=20 members of your family? Or was there a club?=20

    KO: Most of this I did alone. I had a brother who was two years = younger and=20 we sometimes experimented together. One thing we did was make a one tube = radio.=20 Before that we made all kinds of crystal radios which were very limited = in their=20 selectivity. You could only get one or two of the local stations. We = made a one=20 tube radio. With this we could buy battery cells for a penny a piece on = sale.=20 And get up to twelve volts. The normal voltage of 200 volts or so was = beyond our=20 budget. But we made a one tube radio. It worked very well. We built it = and=20 rebuilt it and rebuilt it again. And learned a lot on the way.=20

    DKA: Where was this, Ken?=20

    KO: I was brought up in Connecticut, outside of Bridgeport. It was an = area=20 where machine tools were built, where you were normally expected to = learn=20 machine shop practice. And I did. But there wasn't much in the way of=20 electronics going on. When I was drafted for World War II, I had the = enormous=20 opportunity to go to electronics school in the Navy. It was a great = school. It=20 lasted a year, or at least eleven months. It was set up by competent = people and=20 they gave an excellent education in electronics. They taught us all the = tricks,=20 manipulating, calculating circuits, the rules of thumb for electronics, = and went=20 through all of the gimmicks and tricks and things one should know about = radio.=20 And then radar, and counter measures. It was the most exciting thing a = young=20 kid, a lot of the electronics could go through.=20

    DKA: How old were you when you went to school?=20

    KO: When I was 18 I went off to the Navy.=20

    DKA: Where was this school?=20

    KO: After a few months at Great Lakes at Boot Camp, I went to = downtown=20 Chicago for two months. I lived in the high school [where] they taught = us the=20 basis of electronics. I'm not sure we even used the word electronics at = that=20 time. The program was set up for the Navy by someone named Captain Eddy = who was,=20 before the war, the one who set up television in Chicago. The Kukla, = Fran and=20 Ollie stories were, I think, one of his projects. We then went out to = Monterey,=20 California for three months, nine months in the middle of San Francisco = Bay on=20 Treasure Island. It was a dream for someone who loved electronics.=20

    DKA: Ken, was this the first time you really had a group of people = around you=20 that you could talk to?=20

    KO: Oh yes, yes.=20

    DKA: What was that like?=20

    TO=20 CONTENTS

  • Radar and Vacuum Tube Innovation

    KO: It was just exciting because everybody was thrilled with what = could be=20 done with electronics. Vacuum tubes were largely just for radio. The = development=20 of radar opened up the use of vacuum tubes in such wonderful ways that = we never=20 conceived of. We'd known what radar was in very general terms. But not = in=20 exactly the detail which how everything was done. And some of the = inventions in=20 radar just thrill me today to think about them.=20

    DKA: Ken when were you drafted? You were drafted because = you...struggled to=20 get into radio electronics?=20

    KO: The Navy gave tests to people ahead of time who wanted to go into = this=20 program. And if you were accepted you were then drafted. But my serial = number=20 always had on the end of it, SV, Selective Volunteer. That is selected = ahead of=20 time and then volunteered. So we went in a special program right away. = And stuck=20 with electronics all the way.=20

    DKA: What were they grooming you to be?=20

    TO=20 CONTENTS

  • Electronic Developments at the Navy

    KO: Oh technician...technician. Their goal was to train enough = sailors to=20 maintain the vast amounts of electronics that were being put on = shipboard. The=20 war demonstrated the importance of electronics. The radar, the radio, = the sonar,=20 the navigation and countermeasures were all very intensive use of = electronics.=20 There also were analog computers for antisubmarine work. And all of = these needed=20 large numbers of technicians. And so they set this massive program up to = generate thousands and thousands of technicians. The result was, I = believe, a=20 major influence on the development of electronics in this country. After = the war=20 when we all went to school, everybody wanted to be in electronics. And = out of=20 that came I think the success this country has had in electronics.=20

    DKA: What happened to you in particular? You went to your school in = Illinois=20 then out to the West Coast, and then what happened?=20

    KO: We went through the usual red tape of sitting in camp and going = on a=20 troop ship. And by then the war was over. And I ended up in China and = was=20 assigned then to an admiral's staff and we lived on one cruiser and then = another, maintaining communications for the admiral. While we were there = we had=20 an opportunity to see China and Korea but we had enormous time for=20 experimenting. In the radio shack where most of the time we were on duty = with=20 nothing to do, the crew was broken into three groups. The card players, = the=20 sleepers and the experimenters. It sometimes rotated. We had access, one = way or=20 another, to all kinds of electronic parts, and we did some fascinating=20 experiments.=20

    DKA: Why do I not have to ask you which group you were in.=20

    KO: [LAUGHS] Then there were things that always had to be done = somewhat=20 illegally. The bureau ships laid down rules on how things were to be = done. But=20 they didn't know what things were like 8,000, 12,000 miles away. And we = had to=20 make things and redo things and get our major projects and...=20

    DKA: What's an example, Ken?=20

    KO: Oh... the communications suddenly, within a harbour, were done on = radios=20 that were made for tanks, and ran on 12 volts. There's no 12 volts on a = ship.=20 Well, we could listen with 12 volts, alright, [but] when you transmit we = need a=20 lot of power. Everytime we went to transmit, we had to turn on a motor = generator=20 set. You couldn't call a technician in to turn it on everytime you're = going to=20 transmit. So we had to make a major switching system to allow an = operator far=20 away to turn on the ______ system automatically. None of this was legal = by the=20 rules but we got the job done.=20

    DKA: You mentioned awhile ago growing up, Ken, [about] machine shop=20 practices. I wondered if that began to pay off in your work in the Navy = and=20 later on.=20

    KO: My father was a machine designer. He said it was okay to go into = radio=20 but it was a business you went into because you loved and therefore you = starved.=20 People in the radio business or electronics then really didn't make a = reasonable=20 living. So he insisted I learn machine shop practice first, and I did = that=20 afternoons when I was in high school. It paid off very well, in the = Navy,=20 because I was the only one who could sharpen a drill and do simple = things like=20 that. [And] when we started Digital I was the closest thing we had to a=20 toolmaker, not a good one, but I made the original tools. We used = cutting sheet=20 metal in making parts. I can at least carry on a conversation with = people today.=20

    DKA: When you were in the Navy and had all this time, were you = prepared when=20 the war ended? Were you ready to function? Were you ready to go to work = in a big=20 way? Or were you still in training?=20

    KO: The plan was obviously to go to college. After the war, we had = developed=20 a lot of self confidence that we could fix anything. Part of it was true = because=20 we were trained in a way. And other parts were very naive. But we didn't = understand. We had to go to college. The government did wonderful things = in=20 encouraging people to go to college with the GI bill of rights. So = almost all of=20 us had the ambition to go right into college.=20

    DKA: The feeling must have been very strong to see all this = development in a=20 field that you loved. And we'd won the war. How did you feel coming out? = Did you=20 feel about opportunity?=20

    TO=20 CONTENTS

  • Electronics After the War

    KO: Oh, yes. We felt a lot of opportunity. We felt electronics was = going to=20 revolutionize industry. We could see so many opportunities for = electronics. It=20 was discouraging to see how slow it picked up. After I had been in = college I=20 thought I'd find a summer job that could be useful. Electronics just was = not=20 finding a place in industry. It took a long time.=20

    DKA: I want to go back to that transition from training when you were = in the=20 Navy to your working. Was going to China the first work? That admiral's = staff,=20 was that the first time that you were deployed?=20

    KO: Yes. Being on shipboard was the first time we really had an = opportunity=20 to practical measure use our training. We probably learned technology = but also=20 learned enormous confidence, maybe way out of line for our skills but = we'd=20 tackle anything. Except once. I remember that the ship's navigator = wanted me to=20 fix this flashlight. And I said, any electronics, yes, but a flashlight, = no.=20 [LAUGHS]=20

    DKA: What were you good at in the Navy? I don't really have a picture = of what=20 you did. I know you were fascinated by all the stuff, but what did you = find=20 yourself doing individually the most?=20

    KO: The job I had in the Navy with a group of two dozen people, was = to keep=20 all the electronics going. And we always had someone on duty. When = transmitters=20 had to be changed, we had to do the changing. When anything went wrong, = we had=20 to fix it. When preventive maintenance was needed, that was our job. = When there=20 was some logistical problem in the operation of _______, that was our = job. So=20 the area of troubleshooting particularly fascinated me because it was an = interested puzzle. Finding out was wrong and fixing it as soon as = possible.=20

    DKA: What did you think of the quality of Navy equipment?=20

    KO: Most of it very good. At that time, now looking back at it, it = was heavy=20 and somewhat crude. But that time it was beautiful, elegant, = magnificent, with a=20 few sad exceptions.=20

    DKA: Was it well organized?=20

    KO: Yes. Very sturdy and very well built. With a few pieces that they = should=20 have left on shore. I was still in school at Treasure Island when the = war ended.=20 I finished soon afterward and after a number of stops on the way, I went = to sea=20 for eight or so months.=20

    DKA: What were you on? Was it a destroyer or battlewagon?=20

    KO: I ended up on a cruiser because that was a cruiser because that = was a=20 flagship. And when the war was over we didn't really think about going = out. We=20 knew we still had a long time to go. People joked about the war lasting = many,=20 many years. [I] never really thought about getting out, you know you're = just=20 there forever almost. When it was time to get out the first idea was to = relax=20 for awhile and then get ready to go to school. But I think most people = really=20 planned to go to school. Some people had no preparation for it at all. = Tried...=20 because the spirit, because of the GI bill, was unique, I think of all = previous=20 wars the drive was to go to school and get education. And I think it = made that=20 generation of, from that war, different from all others. People really = drove to=20 go to school. I went to work at General Electric which was in town, = where they=20 were making radios. We made very expensive FM radios. I was a = troubleshooter on=20 the end of the line, [doing] mass production troubleshooting. I loved = it, it was=20 fun. This was in '46. My folks were still in Connecticut. After they = mustered me=20 out I went home and spent a month or two going to the beach and then = went to=20 work for General Electric who were hiring anybody they could get who = could do=20 work in electronics. They had a whole war time of backlog to catch up = with. They=20 were making very expensive FM radios. I enjoyed it and learned a lot. At = the=20 same time I was studying for a college entrance exam. I applied only to = one=20 place, MIT, took the exam and was accepted.=20

    DKA: Why did you only apply one place, Ken?=20

    KO: There was only one place I wanted to go. My grades weren't all = that great=20 but I... [PAUSE] didn't think far enough ahead to think of an alternate. =

    DKA: What did you know about MIT to have made you want to go?=20

    KO: It had a good reputation in electronics from the war and I = suppose that=20 was all.=20

    DKA: And what was it like to be at MIT as a student compared to = today?=20

    KO: We were a extra class brought in February. Largely of veterans so = we fit=20 right in. And we kind of ran the place.=20

    DKA: What does that mean?=20

    TO=20 CONTENTS

  • The M.I.T. Experience

    KO: Well there was nobody who was young and scared, you know. Very = serious in=20 a way and yet very carefree also. I went there in February of '47. And = then went=20 four terms straight without vacation. The first two years were a = standard course=20 for everybody. Interestingly, when it came time to select the major, = many people=20 went into electronics, wanted to go into electronics because they had = exposure=20 to it during the war. MIT set about to interview everyone with a goal of = talking=20 a certain percentage of them out of electronics, because there really = wasn't=20 room. They had calculated that there wasn't going to be that much future = for=20 electronics, so they tried to talk people into going into chemical = engineering.=20 I remember the questions in the interview, and I remember the professor = who=20 interviewed me. I was accepted to stay in; course six of electrical = engineering.=20 Electrical engineering was particularly useful for the future, our = future in=20 computers because many of the people who laid out the course of study = came from=20 the radiation lab and the war time experience. One of the things they = clearly=20 said they had in mind was that if another war started, there was going = to be=20 somebody able to design magnets. Because when they tried to use the = magnetron in=20 World War II, they couldn't find anybody to design magnets. So, out of = these=20 classes came a large number of people who had some simple basic theory = of=20 magnets. And out of that came very useful knowledge for us in the = development of=20 the core memory, the key part of computers. People who came from other = schools=20 without that background were at a serious disadvantage.=20

    DKA: Did you study from the radiation laboratory textbook series that = became=20 so famous, Ken?=20

    KO: No, I don't think there were ever classes. There may have been. = They did=20 develop a number of classes, though, where you could see the influence = of the=20 people who wrote the rad lab series and who took part in their work. One = interesting experience, when we studied circuits, probably the first = course in=20 electrical engineering, I had learned all the tricks and all the ways of = doing=20 fast computation. And manipulation in...in one's head in the Navy. We = could do=20 X, Y transformations and do parallel series networks of components and = do it=20 very quickly. And for several weeks [in] the MIT course, I could do all = the=20 answers in my head. Suddenly it got beyond what I could do in my head. = And with=20 a panic I had to go back and learn the systematic approach to it that I = had just=20 lost in those weeks. Fortunately I caught up and I was able to do it. = But I=20 almost lost out completely because I had learned so much before I got = there. But=20 I didn't learn the systematic approach that would take you to the really = complex=20 questions. There was quite a bit of laboratory work at MIT at this time, = very=20 much so in physics and in chemistry. The most memorable laboratory which = I feel=20 badly has disappeared is the motor laboratory. They had a very large = room, very=20 high ceiling. Quite dusty and dark. Large generators and motors. Oh, six = feet in=20 diameter that you'd set up and run and learn by the sound of them and by = the=20 sound of the sparks and the dramatic result of making a mistake and the = wire=20 would evaporate. A feeling for electricity that too many students missed = today=20 because it's all simulated on a computer. It's not quite the same as = hearing a=20 motor run away and about to explode if you don't dive for the switch and = turn it=20 off. So we did learn some things there that are missing in today's = education, I=20 think.=20

    DKA: Are you the kind of guy that liked to spend lots of time in the = lab at=20 this period?=20

    KO: Oh, I enjoyed it. The electrical engineering laboratory = experiments took=20 a lot of time. Many nights we literally stayed all through the night = working on=20 the experiment, writing up the experiments. So you never went and did = any more=20 of those than you had to because they were so time consuming.=20

    DKA: I haven't heard anything about what you were getting interested = in.=20

    KO: [My interest] was still in electronics. Oh, I didn't know what a = computer=20 was. One of the fascinating things was electronics were medicine and = healthcare.=20 Also machine tool control. Being brought up in the machine tool = industry, the=20 control of machine tools with electronics was interesting.=20

    DKA: It's numerical control.=20

    KO: Numerical control, the like. MIT was making a numerical control = milling=20 machine. When I was out looking for a job, I went to the head of that = project=20 who later became head of the department and I said, I'd like to go to = work here.=20 And he said, "Well, we don't have a contract. We don't have any money, = so we=20 can't hire you." About then I got an invitation to come join the = computer=20 laboratory. They hired only the top ten percent of the class to the = computer=20 laboratory which I think I wasn't quite there. But my love for = electronics had=20 caught the imagination of one of the professors and he recommended me = even=20 though I wasn't the top ten percent. And I accepted. And a day later = Gordon=20 Brown, head of the numerical control machine was running down the hall = and=20 saying, "Ken, Ken, we got money." I said, it's too late. I've already = got a job.=20

    DKA: Probably the most lucky lack of funds in the history of = computers.=20

    KO: Yeah. At least for me.=20

    DKA: How'd you feel to be able to go? What did you know about the = computer=20 lab?=20

    KO: Oh, I knew nothing. Nothing at all. It was classified. It was = military.=20 So no one knew what was going on inside.=20

    DKA: Did you know it was Navy at that time? Was it still Navy?=20

    KO:It was Navy but nobody outside knew what it was except those who = had legal=20 access to it. I was graduating and going on to graduate school and = needed a job=20 as a research assistant. That was 1950.=20

    TO=20 CONTENTS

  • Undergraduate Studies

    I'll make a few more = comments back=20 in the undergraduate area. Some of the things that I found particularly = exciting=20 and fascinating in the undergraduate study is the technology or the = techniques=20 for testing, proving or learning things. In physics they would say, = suppose you=20 looked at an area like a pill box and it got skinnier and skinnier and = skinnier.=20 What would happen? Or in the shape of a tube and it got skinnier and = skinnier=20 and skinnier, what would happen? Or suppose something went to zero, what = would=20 happen to the rest of it? Or went to infinity, what happened to the rest = of it?=20 And looking at problems this way. It is ingrained in you. [It] is = surprisingly=20 useful all through engineering and also in business. That part of the=20 undergraduate training is still one of the most fascinating and one of = the most=20 useful mathematics. A lot [of] time was spent in complex mathematics. = But the=20 key part of mathematics is the simple concepts of calculus where one = looks at=20 everything in terms of slopes, simple derivatives. It gives insights = into so=20 much phenomena, including business phenomena. You never talk of business = in=20 terms of calculus, but it sure is a handy way to look at it because the = balance=20 sheet is an inner glow of the P&L statement. You really have to look = at the=20 derivative of the P&L statement in order to gain knowledge. And = sometimes a=20 second derivative. All of this helps in looking at phenomena we work = with every=20 day outside of the academic world.=20

    DKA: Are you [saying] that calculus gives you a sense of trends in = business?=20 I'm not sure I get the parallel.=20

    KO:If the company is growing, you look at the derivative of the = P&L=20 statement. There are factors there that, for example, show that it costs = to=20 grow. You stop growing, the cost disappears. And the usefulness is, for = example,=20 if you stay at a constant growth, that cost should stay constant. But if = that=20 rate of growth is changing you have to take another derivative. And = there's=20 another cost which should go up and down. Looking at it that way you can = draw=20 conclusions, simple conclusions on a financial statement which are not=20 immediately obvious unless you think of it in those terms.=20

    DKA: So it's a kind of discipline of thinking?=20

    KO: It's just mathematical tricks at looking at things. Now that's of = probably no general interest but... There are a number of things that = are taught=20 in engineering and science that sometimes aren't taught very much today = and I=20 think are lacking. One thing still taught in many places that's still = very=20 important is that every engineer keeps an engineering notebook. = Everything you=20 do, everything you learn, everything you run into, even if it's a phone = number=20 or a piece of data you collected, you write in a notebook. You never = correct it.=20 If it's wrong you cross it over and do it again but leave record of what = you had=20 there. Everything is kept there permanently. It may be written poorly, = but at=20 least it's there. Then there is the belief that [there is] absolute = honesty in=20 what you're doing. If a piece of data doesn't come out right you flag it = as it=20 didn't come out right. But you never adjust it so it looks right. In any = experiment, you may have mistakes in it. But no one would ever waiver or = even,=20 the thought of having any dishonest is...the tradition.=20

    KO:Part of the ideas of science were traditional [and] came from many = years=20 past and from reading and being exposed to scientists. I bought a = recording=20 barometer and a recording thermometer a few years ago, and I felt this=20 overwhelming obligation to keep all the records of temperature and = pressure=20 perfectly dated and filed. For no reason. It's just that if you're = taking the=20 data, you're just supposed to do that. I knew it was foolishness because = there=20 was never no need for me to have it. But it's just part of that = tradition. This=20 comes about from reading of scientists and how things are done. The = other idea=20 that I developed before MIT and afterward is the parallel between = Christianity=20 and science. The books written before those years often were about the = conflict=20 of Christianity and science. But it's obvious that the main theme of = both is the=20 same, which is searching for the truth, which implies a certain = humility. This=20 has turned out lately to be a very important idea, I believe. The = traditions of=20 the church were never absolutes, but searching for the truth. The = scientist says=20 there are very few absolutes but searching for the truth. Today the = scientist,=20 because he's on some kick for ecology or something, will talk about = absolutes he=20 knows nothing about. And we've been exposed to so many people in the = church who=20 are absolutely sure about things they know nothing about. It's = interesting to=20 see how in both cases we've deviated from the original traditions. Some = of the=20 people in the church today would say by the way they act. St. Paul and = Christ=20 were of course, in ignorant times, but we know. And scientists today = will say,=20 "All those old scientists who always cautioned that they didn't have the = final=20 truth." That was the old times. Now we know. It happens all the time = today.=20

    DKA: I guess I'm curious that you talk much about scientist and not = about=20 engineers in your training. Do you see a distinction there, Ken?=20

    KO: They blend together. Engineers should follow the tradition of the = scientists. The engineer is the more practical scientist based on=20 data....[NOISE]=20

    KO: Very few people do much philosophical thinking after they leave = school.=20 When you're in school you think, when I have time, I'll think about = these=20 things. When I was ready to graduate and I was accepted into graduate = school, I=20 had to find a job as a research assistant.=20

    DKA: You were going to graduate school at MIT?=20

    KO: At MIT. Do you remember [in] "Fiddler on the Roof," they asked = "Why do=20 you stay in this town?". He says, "It's home." That's part of it. = Interestingly,=20 the class in those days didn't like MIT. Afterward they did, but I never = thought=20 of leaving and go[ing] anywhere else.=20

    DKA: What do you remember about learning about what was going on in = the=20 computer laboratory? What had you heard?=20

    KO:We heard nothing about the computer laboratory. No one knew = anything about=20 it that I had any contact with. Students tend to know a lot about what's = going=20 on. We had the equivalent of hackers who had access to the telephone = system.=20 People had access to all sorts of things. We knew nothing about the = computer=20 laboratory. There was nothing particularly military going on there. = There was no=20 real project except make a computer. But we knew nothing about it = because the=20 security was that good.=20

    TO=20 CONTENTS

  • Description of the Whirlwind Computer

    DKA: Was there any doubt in your mind what the computer was?=20

    KO:Oh, I had no idea what it was. The concept of the computer, I had = no idea.=20 Entering the laboratory was a little bit like going into a religious = order as a=20 neophyte. They had an attitude about reliability and how you build = electronics=20 which they believed religiously. You had to follow the rules. It was = almost a=20 fixed procedure that you had faith in rather than something you knew = worked. The=20 problem they had is very obvious. They're building a ten thousand vacuum = tube=20 computer with vacuum tubes that had a design life of 500 hours. So, if = you do a=20 simple arithmetic, you could easily conclude the thing should never work = under=20 any circumstances. Very special care had to be taken to make things = work. The=20 vacuum tubes were never turned on. They were slowly turned on. They were = never=20 turned off then. The designs were done with utmost care. Everything was = tested=20 and wide margins were held in every component so that anything could = vary and=20 the thing would still work. Then to top it off they built this whole = computer=20 with ten thousand vacuum tubes in a room, oh, twice as big as this one, = [THE=20 INTERVIEW SITE IS 2500 SQUARE FEET] with long racks. The racks are 22 = inches=20 wide, 11 feet high. Each rack was a digit, so there were 16 of them plus = a=20 couple on the end, filled with vacuum tubes. Every vacuum tube had one = grid, the=20 screen grid, which was brought back to a large telephone switch. So any = tube or=20 any collection of tubes could have that screen grid voltage varied until = the=20 computer failed. You could tell in that grouping if anything was = deteriorating.=20 That way the beginning of the day you could replace any tubes that were = turning=20 weak. The utmost care was in reliability. That was one of the secrets = [of]=20 Whirlwind, and it was one of Whirlwind's contributions to the world of = computing=20 with that extreme care for reliability.=20

    DKA: So you got in the Whirlwind project and you began to see the = whole=20 project. What was your feeling about recognizing what they were = attempting to=20 do? Was this brand new to you?=20

    KO: Oh, it was brand new. I was awestruck. And I loved it. Anybody = who put=20 10,000 vacuum tubes together, I used to tell my wife it was a bridge = builders=20 personality that would build anything that big. Imagine 10,000 tubes in = a room.=20

    DKA: Had you ever seen anything that big electronically before?=20

    KO: Oh, no, you see, during the war a radar set with 150 tubes was = amazing.=20 150 tubes was just out of this world. You couldn't conceive of that. = That=20 building, what with 10,000 tubes, that just was way out. Now a large = part of it=20 was war surplus equipment. Many of the tubes were war surplus. One way = of=20 getting large numbers of things at an inexpensive price. The history of=20 Whirlwind goes something like this. I joined there in 1950. It started = during=20 World War II. The government wanted a computer to run a wind tunnel. = They=20 started off, [while] the war was still going, to make an analog = computer. I have=20 one of the parts of the analog computer in my office. Analog comput[ing] = was too=20 slow and they built a serial digital computer. In time it was discovered = it was=20 too slow. So they built a parallel computer, which is what Whirlwind = ended up to=20 be. It had to be exceedingly fast. Whirlwind had just [a] 16-bit word = length,=20 which in those days was considered ridiculously small. But they were = working on=20 physical phenomena which 16 bits described as very well. People doing=20 mathematics had 48-, 64-bit word length. But the circuits were made very = fast.=20 Exceedingly fast. Each circuit was intensively engineered and put = together in=20 very simple logic to maintain the speed. The thoroughness to the = engineer and=20 the insight into how he did use vacuum tubes, I found fascinating. I = found out=20 who the smartest designer was and sat on his side and learned from him=20 immediately. It was Dick Best, now at Digital. [Dick Best] would design = things=20 in a way that everything was tested, and then draft in a way that = anybody could=20 see how it worked. Instead of having drawings made, all the data is = there and=20 then forget it, he would study it until it was in the form that anybody = could=20 look at it and figure out how it worked. Exceedingly objective. One of = the=20 problems we had in those days was that there was no oscilloscopes to = look at=20 high speed circuits. Tektronix was not yet in business. We had war = surplus,=20 Sylvania cyncroscopes which were quite a large box with almost nothing = in the=20 box. The five inch cathode ray tube, and in order to use it we cut a = hole in the=20 top. Every engineer did his own. I learned from Dick Best how to do it. = Cut a=20 hole in the top and put the wire right into the deflection plate with no = amplifier. The deflection plates had sensitivity as 60 or 70 volts per = inch. So=20 if you were looking at 10 volt pulses, you would see about a sixth of an = inch=20 high pulse on the screen. You also had problems, if you used wire to do = this it=20 would ring and you'd get all sorts of spurious signals. But with that we = developed a technique, or I probably learned it from others, where we = could=20 measure a fraction of volt and amplitude even though the picture was = that small=20 because it was all technique for making measurements, and there being no = amplifiers available to make a bigger picture.=20

    TO=20 CONTENTS

  • History of the Whirlwind

    Dick [Best] was a = great=20 designer of circuits. When he was working I'd look over his shoulder and = learn=20 from him. Others were logic designers and there was something to learn = from=20 [them]. The difference between Whirlwind and the other computers at the = time,=20 there were about three or four being built, they were all six months = from=20 completion. Every year they were six months from completion. All three = or four=20 of them. There's a story which my friends at Harvard say was apocryphal. = But the=20 story says that, Professor Aiken at that time said, "When all the = computers then=20 being built were complete, they take care of all the computation needed = in the=20 world." And he, or whoever said that, was, of course, looking at the = computation=20 then being done and divided that into the capability of a computer and = [it]=20 look[ed] like you didn't need very many. Nowadays we have thousands of = times=20 more computing on a desk and we're still looking for more because as we = get it=20 we find more uses for it.=20

    TO=20 CONTENTS

  • The Basis for Computer Design in the = 1950's

    [The logic=20 for] the computers then being built was designed by using boolean = algebra. It=20 was the mathematics used to design large networks and switches. And it = was=20 directly _______ to computer type work. However the people doing = Whirlwind, Bob=20 Everett, probably being the leader, approached it differently. They = approached=20 design of mathematics and the logic of the computer as if it were a = puzzle,=20 considering all possibilities and picking out the best ones. The result = was=20 exceedingly simple, elegantly simple way of building a computer. The = boolean=20 algebra people ended up with very complex computers. They had simple = circuits=20 and complex logic. Bob Everett's approach was very complex, thoroughly=20 engineered circuits, but very simple logic and very fast. I like to = think that=20 we, Digital, were missionaries to the world to convince them that the = MIT way=20 was the best way. Because the MIT way is what's commonly accepted today. = And I=20 like to think we helped sell the idea. But it definitely was unique. = When we=20 started Digital, we sold modules, little blocks of logic that people = could use.=20 People would ask us how many levels of logic we could do in the speed we = said we=20 could do arithmetic. We couldn't answer that because we followed the MIT = tradition where you did it all in one step. Everybody else did it in = several=20 steps and therefore it was a lot slower.=20

    DKA: You mentioned that sometime in this period you got married. Do = you want=20 to say a word about that?=20

    KO: While I was still an undergraduate, before my senior year, the = neighbor=20 next door to my parents had a Finnish girl visit them for the summer who = was a=20 student in this country. I didn't make out well with her at all. She = went back=20 to Finland and the summer between undergraduate school and graduate = school, I=20 got a job in the ball bearing [factory] in Goteberg, Sweden, as an = excuse to go=20 to Finland and see how I would do with her. So that summer I got there = and in=20 two weeks, became engaged to her. My approach to things often is [to] be = somewhat systematic. I kept a notebook of all the things I wanted in a = wife for=20 a number of years. Every time I heard a preacher preach on the subject, = when=20 preachers used to preach on practical things, you see, not like today, = or=20 anybody would have a lecture, or any ideas [would] come, I'd write down = what I=20 want[ed] in a wife. And when I found her, she kicked off perfectly on = every=20 point; I knew she was the one. She didn't agree that quickly. She was a = Finn,=20 [a] very nationalistic Finn. Leaving Finland was a just strange idea. = When I=20 went there that summer, her brother wouldn't shake my hand because Finns = are=20 Finns and they don't leave. But she agreed to marry me. Then we had a = terrible=20 time getting her into this country because the rules were very strict. = We=20 couldn't get her in as a tourist or as a student or any other way. The = Iron=20 Curtain was expected to come down and then close. The Korean War was on = and I=20 had a friend from MIT who went over to Europe with me but couldn't get = his wife=20 out of Poland and [was in the] terrible position of being engaged [with] = no hope=20 of ever getting his wife out. So I went home, engaged, with a concern = whether=20 I'd get my fiance out. I went back in December at Christmas vacation, = and stayed=20 six weeks til I got permission to bring her in. And got behind in = school. [But]=20 with the generosity and patience of MIT I kept my job there. And with = her=20 patience, [I] got the work done at MIT. It was an important part of = getting=20 anything done I got done.=20

    DKA: Is it true that the term Whirlwind came from your romantic = affair and=20 that was applied to..[LAUGHTER]=20

    TO=20 CONTENTS

  • The Whirlwind Computer - Origins of the Name

    KO: Oh, no, see Whirlwind was well underway by the time I showed up. = The=20 story, before my time, was that its military code name was "tricycle" or = "kiddy=20 car." Something trivial like that. Jay Forrester, the boss, one day came = by and=20 said, "That name has to go. From now on it's Whirlwind." So that's the = only=20 story I know. And it was a good name for it.=20

    DKA: Your fiance must have been amazed. Did she know you were coming = over to=20 work on?=20

    KO: Oh, we exchanged mail so she met me there, yes. We bicycled = around=20 Finland together.=20

    DKA: She must have known you were serious if you'd come across the = world for=20 her.=20

    KO: Yes, yes. She never went back to Finland for 16 years. People = can't=20 understand why, [but] if you saw the film, "Dr. Zhivago" you realize = that in war=20 time you had idea[s] never to leave your family. Once she had a new = family she=20 wasn't going to leave. Once she got back she saw the world was different = and she=20 went back every year to see her mother. But when the memory was of = wartime=20 Finland, she wasn't going to take any chances with the military or = immigration=20 people of ever getting separated again.=20

    DKA: Back to work in the lab, what was your work? What were you doing = yourself?=20

    KO: First I was doing small projects. My first job was to make a = digital to=20 analog converter to drive the cathode ray displays. This sort of thing = we take=20 for granted today on our personal computer[s] where we have pictures on = a=20 cathode ray tube. At that time [it] was a rather unusual idea, one of = the=20 developments that came out of MIT and influenced the world of computing. = Converting the signal from a Digital number to an analog voltage to = drive the=20 cathode ray tube was one of the devices I'd had. My contribution was to, = here=20 was a clever little circuit that used very few vacuum tubes and made a = precise=20 unit. This circuit then became a very important part of the core memory. =

    TO=20 CONTENTS

  • Core Memory

    KO: The core memory was invented by Jay Forrester, a brilliant idea. = The core=20 memory idea was not the idea of storing information in a core. That had = been=20 done before. The clever idea of Jay Forrester's invention was the way of = selecting the core. So you could put thousands of cores together and = select them=20 very quickly and easily. It still took a large number of current sources = to=20 drive it. The number of current sources stifled the future development. = No one=20 had the nerve to build that many current sources. The [original] memory = for the=20 Whirlwind computer was a storage tube. The storage tube was quite large = with a=20 neck coming out of it about 8 inches in diameter. It was really a = cathode ray=20 tube with two guns coming into it. Stored on the face, on the form of = dots on=20 this, were the ones and zeros. And when it was working well, (and it was = hard to=20 get 32 of them ever working together, all you got was 256 bits per tube = which=20 meant when the whole thing was working), you had 512 bit words of memory = which=20 is a joke today, [and] even then was marginal. It took great cleverness = to solve=20 problems without much memory. The pressure was really developing more = efficient=20 memory. Jay Forrester came up with this brilliant idea of using cores. = His first=20 idea was a gas discharge where you have a big bottle of neon and near, = crossed=20 wires a spark would develop. Would stay there for one and would = disappear for=20 zero. This idea then was to put a core at each intersection. And use = that for=20 memory. And this progressed as great excitement.=20

    TO=20 CONTENTS

  • The M.I.T. Lab

    KO: The laboratory had some very productive ideas, or ideas that made = the=20 laboratory very productive. There was a lot of trust, a lot of freedom, = a lot of=20 competition between very bright people. But a lot of openness. And with = that=20 communication was free. And we all had to write a report every two = weeks, maybe=20 only one paragraph or a few sentences. Even if we did nothing we had to = write=20 down that we did nothing. With that the communication was very open. = When you=20 had an idea you immediately had everybody knowing about it. And if it = was a good=20 idea you had support. And if it was a bad idea you quickly realized you = should=20 just quietly go away with it. So with the idea of how to do a memory, = there was=20 no problem of getting it across. We started a single plain memory, 16 by = 16=20 cores. Each core was a ceramic bobbin, about a quarter of an inch in = diameter,=20 maybe an eighth of an inch diameter inside. Wound on that was a very = thin foil=20 of magnetic material and then heat treated to make sure there were no = tensions=20 in it and that was the core. There were four wires strung and then in = order to=20 drive it, it took 16 current sources times four. Two for each coordinate = and=20 that same thing again for reading and writing. The idea worked but it = still had=20 so many vacuum tubes in it scared people. People still didn't feel bold = enough=20 to build a big core memory. I had an idea for using a magnetic switch to = eliminate all the current sources. Get rid of all the tubes and drive = them with=20 magnetic switch or drive them with cores which were selected the same = way memory=20 cores were selected. So when I quick changed my thesis, somebody else = finished=20 the original one. And it worked. It never was a great success. It never = really=20 contributed an enormous amount to anything except for one thing. It was = a great=20 academic interest. At least to me. But what it did do was, for awhile, = got rid=20 of the hangup people had about having too many tubes. So there was a = spurt of=20 interest in core memories. It solved the problem in people's heads. They = went=20 off and built it with vacuum tubes. But the contribution of the core = switch was=20 it got people's hangups to disappear long enough to get enthusiasm. The = story=20 [that] goes parallel with that is (and I didn't know about this until = just=20 lately), IBM's next generation of computers use core memory ________ = switches.=20 But I never knew that. The problem we had with the core memory, (the = first one=20 was going to be 16 by 16 or 256 words, and 16 digits long), was how to = test it.=20 We understood from experience that you really had to test every possible = combination. Because [no matter] how thoroughly you designed things = there's=20 always something that might go wrong or some combination of things that = might be=20 wrong. And people were not about to trust the core memory unless it was = truly=20 tested in an environment that was tested. So we set about to build what = we=20 called a memory test computer. It was supposed to be an honest to = goodness=20 computer that would really run and test the memory, but not a computer = that=20 designed to be useful. I was given the job of building the computer just = as soon=20 as my thesis was done. I think I was still a graduate student and it = cost a=20 million dollars. I can remember being impressed of how much a million = dollars=20 was. How much work it took to spend a million dollars. Now I'm impressed = at how=20 little effort it takes to spend a million dollars. So we built a 16-bit = machine.=20 My way of showing off was to build it in a room in a straight row of = racks with=20 a console in front of it, with enough room for the photographer to stand = back=20 and take pictures of it. We naively showed off by saying, look how easy = it is.=20 That's kind of the young academic approach. The problem with that was = that=20 people believed it was that easy and never took it seriously. We learned = later,=20 for the next machines I was responsible for, to do it with a little more = flair=20 than that. We made a homemade wooden console with cabinets from the = local=20 distributor. Afterward we learned to put color in it.=20

    TO=20 CONTENTS

  • Listening to the Computer Late at Night

    KO: The machine ran well. The first night it ran, my wife was out of = town.=20 And we stayed late at the lab and it finally worked. Everybody else went = home=20 and I stayed there and listened to it work. We put a loud speaker on = every=20 computer we built because you always wanted to be able to play music or = make it=20 do things. So I had the computer on the loud speaker and as long as the = tone was=20 constant I knew it was working. So I went in the ladies room and laid = down on=20 the sofa with the door open and fell asleep with my ear tuned to that = sound so I=20 knew that it went all night long without a glitch and that was a = significant=20 test. As soon as the machine was truly completed and within one day of = it being=20 working, the people in charge made the decision to shut down the storage = tube=20 lab and switch everything over to the core memory. The second memory was = started=20 immediately. The first one, we painted everything. The second one was = bare=20 aluminum because we weren't going to take the time. So we had two = memories=20 there. Whirlwind, the memory test computer was there without a memory. = The=20 machine we worked so hard on suddenly, instantly was sitting there with = no=20 memory. And therefore quite useless for awhile.=20

    TO=20 CONTENTS

  • Modular Computer Design

    KO: Today we tend to build computers by putting them on a small board = and=20 designing everything right there. But an important part of the early = development=20 of computers was to make them modular so that things that were used many = times=20 were made identical and used over and over again. It wasn't always = obvious. Not=20 everybody agreed with it that way, but one of the first modular = approaches was=20 Whirlwind. This is one of panels of Whirlwind. [HOLDING RACK FROM = WHIRLWIND=20 COMPUTER] There were probably 16 of these all in a row, all identical. = The=20 original theory was, if one went bad, you took it out, replaced it. At = Whirlwind=20 it never quite worked that way because it was easier to troubleshoot in = place=20 because you could reach behind, take out a vacuum tube. Those tubes had = the=20 advantage because they didn't fall out easy, but they didn't come out = easy=20 either. You could change a part and troubleshoot without turning off the = power.=20 There was 250 volts positive, 150 volt negative. You could learn by = touching it,=20 approximately [what] the voltage was. But shutting it down was such an = operation=20 you never thought of shutting it down. You did things carefully, you = never=20 thought of hurting yourself, but [because] you could ruin something if = you did=20 it poorly. The parts today looked antique. The capacitors were mica with = foil=20 between them embedded in plexiglass. The resistors were carbon and = normally not=20 very precise. The terminals were, for some reason I can never explain,=20 silver-plated. It sounded quality. But they corroded, were impossible to = solder=20 and created all kinds of problems. But that was the tradition of the = day. The=20 Whirlwind was made up of these modules. I don't know what this one did. = This one=20 was a program counter. This was one digit of a program counter. So, as = your=20 program went step by step, it kept track of what the last step was and = set up=20 the next step. I think there were twelve digits of the program counter = so there=20 are twelve identical units. They're put in a row and they would keep = track of=20 how many steps the program went through. Supposedly there was a spare = and you=20 could have swapped it. It just happened to never to be done. The next = step in=20 modules development at MIT was to make true modules where there was a = large=20 number of the same thing and they look something like this one where you = not=20 only made it easy to swap units that were defective, but you also gained = density=20 in the third dimension so that you could get a lot more stuff in a much = smaller=20 area. This was a module used for building MTC computer out of...=20

    DKA: How many tubes did the MTC have?=20

    KO: Thousands of tubes.=20

    DKA: And this was the reason why because you had thousands of these=20 components?=20

    KO: Yes. So there was hundreds of these components. Most of them had = just two=20 tubes in them. With transistors it was much easier because with these, = each tube=20 had approximately five, ten watts of filament power. That meant that = every time=20 you built something, you had huge transformers just to drive the = filaments. The=20 voltage on the plates were 250 volts. Any current of 250 volts had a lot = of=20 power. With transistors, life became easy. The first transistor computer = we=20 built at MIT was the TX-0 computer. I didn't have faith in putting = [transistors]=20 in circuit boards. Maybe poor judgement on my part. But we built them = into a=20 tube like this. [HOLDS BOTTLE FROM TX-0] There was one transistor. They = were fit=20 in a small socket. Then you could have a high density, and they had = color codes=20 on the top so that if a red one went bad, you'd just put another red one = in.=20 These transistors were made by Philco. They were the last of the high = speed=20 transistors. No one else could make high speed transistors. So we = grabbed hold=20 of and designed circuits to match their characteristics. They were so = delicate=20 that if you combed your hair and touched one, you burned it out. It took = a=20 special set of circuits to do it. They cost several dollars each. The = next=20 computer module we built was the TX-2. We had printed circuit boards. = [HOLDS=20 TX-2 MODULE] This module we designed for high density. We had real = solid, secure=20 sockets we thought at the time. The printer circuit boards were here in = special=20 metal size. It looked quite attractive. It also was color coded with = colors on a=20 handle. The transistors are through that hole there and other components = are=20 laid out and you can see the components become more moderate now. = They're more=20 compact. They look a little more professional. The connector is solid = and rugged=20 and shows our lack of faith in connectors at the time.=20

    DKA: Ken, is this again something that you designed all the circuitry = on?=20

    TO=20 CONTENTS

  • Making Test Equipment

    KO: In general [the design of] the circuits involved many people. The = idea=20 [of] these circuits, I probably did when I was all alone on a project. = But in=20 time many people got involved. One of the ideas they had at MIT before I = showed=20 up was to make what they called test equipment. It was a set of modules = that=20 would do single operations like a flip flop, a gate, a delay unit. With = those=20 you put on your experimental bench, develop an environment for testing a = circuit. That detail[ed] testing of circuits was possible because of = testing=20 equipment. When we started Digital, the first product we had was the = equivalent=20 with transistor computers and this is one of those. [HOLDS DEC LAB = MODULE] You=20 could arrange these in a rack and do specific operations at a speed that = no one=20 else could accomplish at the time. But you would wire them together with = these=20 pig tails and make a counter, set up pulses or anything you want to do = for=20 testing. That was our first product. It allowed the rest of the world to = design=20 the logic they needed for military projects and other computer projects. = When we=20 wanted to make something permanent like a computer, our first computer, = we took=20 the same circuits and put them into a frame which could be stacked quite = densely=20 without the cover. These we called our system modules. They were a key = part of=20 the business for many years. The transistors had changed by this time. = The=20 components had gotten smaller. We had a unique idea in driving these. We = said we=20 would make a simple rugged power supply and design the circuits to = tolerate the=20 variations in the power supply. We also used diodes to generate three = volts, the=20 base voltage right on a board which meant this was almost completely = tolerant of=20 noise on the power line, lightning or anything else that wiped out = computers at=20 the time. That made our computers very rugged. When some of the space = programs,=20 everybody's computer was down. Ours was still running because we = generated our=20 base voltage right on the board.=20

    DKA: You were talking about what service that component provided. = Could you=20 go through a bit of that again.=20

    KO: We had a box for every circuit you'd want to build a computer or = computer=20 environment. This one happened to be a pulse amplifier. If you look at = the=20 diagram, there's a gate and an amplifier. So, if you put in two signals = here and=20 they were in the right combination, you'd get a pulse out here at a = standardized=20 size. Other boxes would be a flipflop that would store information. A = complex=20 set of gates which would allow you to do logic and delay units that = would allow=20 you to accomplish other activities. With that set of pieces, people = could do=20 almost anything they want to do at high speed.=20

    DKA: The degree to which you did modular thinking, was that unusual = in this=20 business at this point?=20

    KO: No, no. By this time it was quite commonly done.=20

    DKA: And what's unusual is the ruggedness with which your components = were=20 designed?=20

    KO: On almost anything someone does in the computer business, you can = go back=20 in the literature and prove someone had done it earlier. In the case of = going in=20 the business selling modules, there was 45 people doing it at that time = we went=20 in business. All doing poorly. Our contribution was first of all we had = the=20 circuits we took from MIT which were fast. No one else could do fast = runs. We=20 also had an interesting business idea. Most of them went into the = customer and=20 said we will charge you ten percent less than what the other guy offered = to sell=20 the same thing for. We went in and said, here's our literature, tell us=20 everything we know about the circuits and here's a fixed price list. And = there's=20 no dickering. With that we changed the industry in a year. Everybody had = a set=20 of literature and everybody had a price list. That was our main = contribution=20 besides speed.=20

    DKA: You had good circuits but you also had a winning strategy.=20

    KO: Yes.=20

    KO: The chronology, as I remember it, this [HOLDING WHIRLWIND MODULE] = was=20 probably designed in '48 or '49, well before I came to Whirlwind. = Probably [at=20 that time] the first module was built as part of the control element of=20 Whirlwind. This module then was designed about 1950, '51, for more or = less high=20 production. The transistor work, the transistor computers we started '55 = or so,=20 '54. We were working on air defense system. I spent a year at IBM = representing=20 MIT. As a rest cure, they allowed me to work on transistors, outside of = the=20 defense part of the business which everybody else was working hard on. I = could=20 have no staff or space. And then these came [HOLDING TX-0 MODULES] = probably '54,=20 '55 and this one '56 [HOLDING TX-2 MODULE]. And then as we started = Digital, this=20 one [HOLDING DEC LAB MODULE] was '57, '58 and then on. We felt quite = confident=20 with the way we were building computers. We knew we could do almost = anything we=20 wanted to do, but the big limitation was the memory. Storage tube = memories=20 weren't reliable, worse than that. And they would never be big or fast. = [When]=20 the idea of the core memory came along, it just raised the possibility = of making=20 real computers. Here's a plane from one of the first ones we built. = [HOLDING=20 WHIRLWIND CORE PLANE] You can see there's a thousand-twenty-four [1024] = cores=20 arranged in a square array. The way it worked was quite simple. The = direction of=20 the magnetization in a core decided whether it was a one or a zero held = there.=20 If you put current on one line it never was enough current to switch the = core=20 from one state to another. But if you put current in two lines, where = the=20 current went from two wires to the same core was enough to switch it = over. With=20 that you could select a core and read a one or zero into it. This meant, = however, 32 drivers on this side and 30 drivers on this side for just = one=20 direction of current. Plus another 32 and another 32. That meant 4 times = 32=20 drivers. You can see those on the memory here. Here's 32 drivers. And = here's 16=20 drivers times 8 around here which ended up being able to drive 32 lines, = 4=20 directions. Each driver just contained one tube and one small tube. That = drove a=20 good half ampere into each wire. This plane comes from one of the first = two core=20 memories built. And right here we have a complete memory which happens = to be the=20 third one. [POINTS OUT WHIRLWIND CORE MEMORY STACK] By the time we made = the=20 third one, the plane's a little smaller. But you can see there was one = plane for=20 every digit of the memory stacked up here. The drivers going in each = direction=20 came from these four sides top and bottom. There's a wire for each of = the digits=20 here. This made a thousand twenty-four [1024] quite reliable core = memory, and=20 this made computing truly possible.=20

    DKA: How many numbers could you store?=20

    TO=20 CONTENTS

  • The Impact of Core Memory

    KO: The 16-bit word, they called two bytes. So there's a thousand = words, or=20 2,000 bytes, which then was a large number. Today a kid with a personal = computer=20 would laugh at you for having that small a memory. The core memory=20 revolutionized computing because it gave the promise of great things. At = the=20 time we never dreamed of large memories because every core has to have = five or=20 six wires put through it. But in time, as the demand grew, the = capability of=20 making them grew. People invented all kinds of machines to do this=20 automatically. But to the very end, the bulk of it by far was done by = girls=20 stringing wires with long needles. We at Digital at that time were a = small part=20 of the computer industry. But at the peak of the core business, we alone = made=20 four billion cores a month. We had girls in Taiwan string each one of = those with=20 five wires through them. Now I have in this little salt shaker some of = the cores=20 we made. [HOLDS SALT SHAKER CONTAINING CORES] When it was full it held a = million=20 cores. They are so small that I can't see the hole in them, even with my = glasses. I think they're about 8-thousandth of an inch in diameter, and = the hole=20 is about 4-thousandth. And they're the size of pepper grains. I can see = the=20 hole, but the idea of putting four wires in there and doing it with 4 = billion a=20 month is just astounding. Here's the specifications for its size and its = chemical mixture. Now with that, people made big memories. And the = computer=20 business spurted forth. The minicomputer business became practical = because we=20 could make inexpensive, very powerful, quite large machines for very = little=20 money. Since then, the semiconductor memory has taken over. It is just = so much=20 easier to use. It's so inexpensive and when you hear about one, four, = eight,=20 sixteen megabyte memories, they're made up of little tiny chips of = ceramic doing=20 the things that we used to do with cores. So as miraculous as these = little cores=20 were, the miracle of the semiconductor memory is much more so. And it = continues=20 to get better every year. The prices go down, the size goes down and = there seems=20 to be no end. The little things that we take for granted, now we have = computers=20 in our automobiles and computers, in our microwave ovens, and computers = in our=20 washing machines. It comes about because the memories now are so cheap. = But a=20 key part in the history of computers was this ceramic core memory. The = first=20 core memory we made, we only made an experimental one plane one. It had = metal=20 cores. It made a very thin foil to make them fast. We knew that it was = possible=20 to get magnetic ceramics called ferrite that would have the = characteristics we=20 wanted, but the builders of ferrites didn't think it was very promising. = One day=20 a company in New Jersey brought some samples made on a washer dye, about = a=20 quarter-inch diameter. We jumped at those. My thesis came out of using = those.=20 And with that small, then very large, about a quarter - we call[ed] them = Cheerios because they're the size of the breakfast cereal ring. From = that came=20 smaller and smaller ones and the small ones I still get stuck to my = hand, are=20 the limit in how small they got. But developing the material was a = different=20 story. MIT set about the traditional scientific way to pick the best = mixture of=20 materials to make these. There were three components to the ferrite; = ferric=20 oxide, ferrous oxide and magnese dioxide I believe. They all cost about = a dollar=20 a pound. But the mixture of the three compounds took a long series of=20 experiments. They tried every combination, plotted them out, fired them, = tested=20 them, and decided what the mixture was. The ceramics company in New = Jersey, an=20 old German ceramacist who by guess or by intuition, or years of = experience,=20 mixed up his first mixture and hit exactly the same spot as all the = research=20 done at MIT. It shows a place for science and there's also a place for=20 intuition.=20

    DKA: Describe this computer for us.=20

    TO=20 CONTENTS

  • The TX-0 Computer

    KO: [STANDING AT THE TX-0 COMPUTER] When I was given the opportunity = to work=20 on a transistor computer, the idea was kind of new, it was exciting and = we had=20 knowledge of the very fast _______ transistor _____ which we had built a = very=20 fast computer. The rules were, I could hire nobody and have no space. I = studied=20 the rules carefully and found all the loopholes. I somehow was able, one = way or=20 another, to get three or four people to work with me. We discovered that = hallway=20 was not space. So we moved my office into the hall and put walls around = it. We=20 then traded that space for a space in the basement which was less = desirable but=20 bigger. With that we were able to do our work. We discovered that part = of the=20 basement in Lincoln Laboratory was nonfinished. It was just dirt. We = talked=20 people into pouring concrete floor there and then we talked people into = putting=20 a light colored floor. When they discovered what we had done they said, = never=20 again. We talked them into twice the light level of anyplace else in the = laboratory. And when they found that out they said, never again. And = then we had=20 the walls with different color. The walls there are just normal military = type. I=20 can't remember if it's beige or green, you know it was just a bland = color. We=20 had a bright color. And then we set about to make a computer that would = attract=20 attention. We discovered with the MTC [MEMORY TEST COMPUTER] computer, = that blah=20 looking computers, never really attract attention. People, you'd think,=20 particularly scientists, would be interested in the specification, the=20 capability. But things have to be colorful to attract attention. So we = set about=20 to make as modern a design as we could. Now it looks quite naive. But = this is=20 it. You know it had rakish lines like race cars were supposed to have = because of=20 the way they took pictures of them. And we picked a color which is just = as=20 opposite from the traditional black wrinkle finish which was World War = II. The=20 modern color used by the laboratory was gray hammertone. It looks so = military=20 and blah. So brown and beige just seem like a dramatic change. And = that's why we=20 picked this color. And we tried to make it look a little modern. About = as modern=20 as we could. The result was when head of the laboratory had visitors he = of=20 course brought it to our laboratory because we set the computer back = from the=20 door for good pictures and showed it off with a little bit of flair. It = was the=20 place they took visitors. Even though we didn't break any rules, we = exploited=20 all the things they didn't have rules on yet and made one that was more=20 exciting. The construction of it is all the things we had learned to put = in=20 computers. There's a loud speaker and amplifier underneath the table for = playing=20 music or anything else you want with the computer. The cathode ray tube = we=20 automatically built into the computer. At that time there were 4,000, I = think=20 4,000 lines because we focused on one spot at a time instead of a raster = like we=20 do today. We use the light pen which is the equivalent of the mouse [or = joy=20 stick] we use today. This is what we use in the aerodefense system. With = that=20 you could draw, play games and all the things you do for the house = today. We=20 used to have a light bulb for every flip flop. We used Japanese model = rail road=20 lamp bulbs. We were joking that we probably confused the industry = watchers over=20 there with that order for lamp bulbs. The model rail road business was = booming.=20 The machine is made of these little tiny bottles and larger modules. It = made it=20 quite easy to make the unit. The transistor was fast but very fragile. = The=20 circuits had to be designed around the transistor. It took 12 = transistors to=20 make a flip flop. I believe it is the design that grew into that = integrated=20 circuits logic which has become very popular. Because they have exactly = the same=20 problem with the power and the transistors as we had there. And so these = circuits, I think, were the basis for the modern computer circuits.=20

    KO: The circuitry in this computer was built around the Philco = surface=20 barrier transistor, a magnificent piece of design for a style transistor = which=20 was just about to become obsolete. It was very expensive but very fast, = and very=20 intolerant of power or spark or discharge of any kind. But we designed = the=20 transistor circuits around this. And we made them very fast and the = circuits, I=20 believe, were the basis for the commonly used T-squared-L [TTL] logic = that=20 people build computers out of today. The reason for building the TX-O = computer,=20 this was about 1955, was to demonstrate how efficient in power, how fast = in=20 speed, and how easy it would be to build a computer for a defense. Now = the=20 project wasn't classified. We published everything. We told everybody. = Had a lot=20 of interchange with the rest of the world. But the goal was clear that = if we had=20 a chance to make defense computers over again, it's obvious that doing = it with=20 transistors would just save so much heat and so much space and be so = much=20 faster. This is a very close to being equivalent to the modern personal=20 computer. Someone sits down in front of the oscilloscope, with a light = pen and=20 plays games, does things, is creative. Word processing wasn't yet = developed.=20 Games weren't yet well developed. But in general it was, you might say, = one of=20 the first personal computers. We designed it as a demonstration but then = people=20 did computing on it. When they had a problem that would lend itself to = this they=20 used it for computing. They'd bring it [THE PROGRAM] in the form of a = paper tape=20 which they generate on a flexowriter. As they typed it, the paper tape = would=20 write the information. In the same way we store things on a disk today, = the high=20 speed photo electric tape reader would read that tape very quickly and = then they=20 would go ahead and do programming.=20

    DKA: When you say this is kind of a personal computer and you talk = about=20 pictures, [CREATING IMAGES ON A CRT WITH A LIGHT PEN] it really confuses = me=20 because it so little resembles the PC and I still don't know.=20

    KO: Oh it's exactly the same as a PC. You see what you see in a PC is = the=20 keyboard. The cathode ray tube and the light pen. So this is indeed a = PC. This=20 [THE TX-0] was designed to be a demonstration of the reliability, the = capability=20 of transistor circuitry. And making a fast, inexpensive low-powered = computer.=20 The unit itself really could do what a personal computer does today = limited only=20 by the fact that the memory was small. You could make pictures on the = cathode=20 ray tube, change them, modify them. Read your program in, take your = program=20 home, play games, do the things you do today on a much smaller scale = because the=20 memory was very small.=20

    DKA: Did you see text on the screen?=20

    KO: Oh yes, the text was very commonly done on the screen. Now as I = remember=20 this was an 18-bit machine. All the machines we had up until then were = 16-bit.=20 It was 18 because we stored a character in 6 bits. Therefore we could in = one=20 word store three very efficiently. So we went to 18 bits just to store=20 characters. When we started Digital we also had 18 bits because we could = store=20 6-bit characters. The world standard later on became 8-bit characters = and we all=20 went to 16-bit or 32-bit computers, interestingly, which was the ones we = had=20 originally. So this is in a very real sense a personal computer. You = could even=20 say that of Whirlwind. It took 2500 square feet. The console was a = walk-in room=20 as big as this loft here. But in a real sense it was a personal computer = and did=20 personal computer things.=20

    DKA: Do you want to talk, Ken, about how some of the ideas and some = of the=20 thoughts and plans that you had when you developed this computer led you = to=20 think about your own computer business?=20

    KO: When this [THE TX-0 COMPUTER] demonstrated the usefulness and = [the ease=20 of making] computers we started a bigger one called TX-2. People often = ask what=20 happened to TX-1. TX-1 was the first one designed and I said, "no way = are we=20 going to build that one. It's too complicated for a first one." So we = built the=20 simplest possible machine which was this one and then skipped TX-1 as a = name and=20 went to TX-2 which was a very large machine. I was building the = hardware.=20 Somebody else was designing the logic and they couldn't settle down. So = after a=20 year or two of that I got impatient and left. That was '57. There [were] = a=20 number of reasons for leaving. One was we [had] published what we had = done,=20 demonstrated that you could make computers very effectively, much better = than=20 anything done with vacuum tubes by far. The commercial world just smiled = at us=20 and said we were just academic. Of course, today, we smile at people at = MIT and=20 say they're just academic. So just showing them it could be done was one = of the=20 reasons for going into business. The things we took from MIT were first = of all,=20 the idea of an interactive computer which was unique. In those days you = dropped=20 your problem in the form of a stack of IBM cards in a slot. It went into = the IBM=20 machine. The next day you got your answer back, and it usually was = [that] you'd=20 made a mistake. With interactive computing, you put a problem and you'd = try=20 something and you [were] instantly told it was a mistake. You could = interact,=20 get things done fast, the things we see in personal computers all the = time. That=20 concept was strange and the idea that that concept should be introduced = in the=20 world. Even more important than that, however, was the demonstration we = had at=20 MIT that where you had a group of people who were bright, wanted to work = hard,=20 but if you showed trust and openness and confidence and let them work = hard, they=20 could turn out amazing work. So the human ideas that came from MIT were = probably=20 the most important. These are the ones we tried to maintain at Digital = where we=20 hire the best and we can hire the best because we have the ideas. And = then trust=20 them, set the general goals so they know where they're supposed to go, = but then=20 give them freedom to be creative. Propose, argue, and then show great = trust and=20 great confidence and they do wonderful things. Those are the ideas we = had at=20 MIT. There's one other reason, too. I always thought that what I wanted = to do=20 was experiment with electronics. I'd gotten to the point where I thought = I could=20 talk people into any project I wanted. That probably wasn't true but I = had that=20 feeling. There was one thing missing which I never thought of before, = and that=20 is nobody cared. It was important to do something they would care about, = so we=20 set about to do something in business that people would care about. And = that's=20 how we started Digital.=20

    DKA: What was it that people didn't care about? And what was it that = made=20 people care about?=20

    KO: We demonstrated all the ideas of high speed transistor computers, = and we=20 thought the world would be waiting in open arms for this. Nobody cared. = And it=20 turns out that it takes more than ideas. You've got to sell your idea. = So we set=20 about to sell the idea. Now there's some lessons there for people. One = of them=20 is it seems like being left alone doing research would be satisfying. = Basically=20 it's not. Unless somebody notices it. And secondly, getting an idea, no = matter=20 how good it is, isn't enough. You've always got to sell the idea. = Putting color=20 into this thing was part of selling the idea. And that's what we set = about to do=20 at Digital. The idea of starting a company was not well developed then. = It was=20 strange. A number of companies had started during the Korean War. A = number were=20 no longer in existence. In 1957, many of them were in trouble. A = recession was=20 starting. The idea was not a popular idea. We were told that the = American=20 Research and the Development Corporation were set up just to do this so = we went=20 to see them. That's the business they were in. But they were worried = because=20 some of their investments hadn't paid off very well. But they were = fascinated=20 enough to listen to our proposal. They told us we could go to their = board of=20 directors and present it and see what happened. They gave us three bits = of=20 advice. One was, don't use the word, "computer." Because FORTUNE = magazine said=20 no one was making money in computers and no one was about to. So we took = that=20 out of our proposal. We were going to make modules first, anyway. And = they said,=20 "don't promise five percent profit." You see we looked in the library. = All good=20 companies seemed to make about five percent on sales. The staff said = that if=20 you're asking someone to give you money, you've got to promise better = results=20 than that. So we promised ten percent. And we made about ten percent = most of our=20 history. If we had looked for five, or aimed at five, we probably never = would=20 have made much more than five. The third thing they said was, "most of = the board=20 is over 80, so promise fast results." So we promised to make a profit in = a year.=20 The other side of the story is that we really did, after 12 months, make = a=20 profit. It was so small you couldn't tell if it was plus or minus. But = it was=20 like $3000 plus. And we brought it down to General George Doriot. [HEAD = OF=20 AMERICAN RESEARCH AND DEVELOPMENT] We dropped the financial statements = on his=20 desk. He looked them over and looked up and scowled at us, which kind of = set us=20 back. He said, "Sorry to see this. No one has succeeded this soon and = ever=20 survived." His lesson of course, was, success is the worst danger in = business=20 and in everything else. Maybe because of his warning we're still here. = But=20 that's also proved to the rest of Digital that success has done more = harm to=20 people than anything else.=20

    DKA: Who is General Doriot?=20

    KO: General Doriot was a Frenchman, came to MIT, but went to Harvard = instead.=20 And became a professor at Harvard Business School, for many, many years, = in the=20 20's. [During] World War II he became a general in the American army = with a=20 terrible French accent so that when I met him I didn't know which army = he was a=20 general in. But he was very popular as a professor. [He] has a strong = following=20 still through the business world, his lessons being very practical. = Integrity,=20 quality, honesty, doing the right thing. He also then became President = of=20 American Research, just to start new companies. His contribution to us = was to=20 encourage us, give us support and show patience and encourage the=20 characteristics which he always taught in his classes.=20

    TO=20 CONTENTS

  • The American Research Company

    KO: American Research was unique in a number of ways, probably all = based on=20 General Doriot. First of all, they were the granddaddy of all risk = capital=20 companies. Since then there have been many risk capital companies. None = of them=20 have accomplished what American Research did. Some of them are = financially more=20 successful, but they never made the contribution. American Research, the = General, had the belief that they made the long-term investment. They = wouldn't=20 buy and sell companies at the first opportunity. They would stick with, = [and]=20 work with, the company until they were successful, or until they failed. = This=20 sounds obvious, but it's very hard for someone who owns a major part of = the=20 stock to be patient. The General really preached this and really = practiced it.=20 It was his contribution. We did well for most of the years. Any other = company=20 would have attempted to sell when somebody was doing well and clean up = on the=20 profit. When things were going poorly, people would be tempted normally = to sell=20 to get rid of the problem. The General was patient then, too. He also = had a lot=20 of simple rules for running a business which are always helpful to keep = in mind.=20 Most of his ideas he didn't present in a way you had to accept. He = presented=20 them in a way which, after it was done, you thought [you had thought] of = them=20 youself. Or if you didn't accept them, there was no hassle. There are a = few=20 exceptions. He said you never want a lawyer on your board, you never = want a=20 banker on your board. These are black and white and you'd have to = definitely go=20 against the General to pick either one of them. He was always there as a = mentor=20 and for help.=20

    TO=20 CONTENTS

  • Digital Equipment Corporation - The Early = Days

    KO: When we started Digital there were three of us. I had asked a = friend from=20 Lincoln [Laboratory] whom I had worked with to work with me on a = proposal. I=20 pretty much had the technology worked out and Harlan Anderson and I = studied=20 history of other companies and the financial statements and laid out a = proforma,=20 financial plan for the company. And then proposed it. When we started, = my=20 brother [Stan Olsen] joined us the first day. We bought the machinery = and=20 started the processes and made the silk screens, and etched the boards = and=20 dipped them in solder. We did everything, the three of us. In time we = hired=20 secretaries and a few other people. We grew quite slowly. American = Research gave=20 us $70,000 and that lasted eight years. The nice thing about $70,000, = you can=20 watch each one of them. We bought things in the hardware store. We were = very=20 cautious and very careful, and learned a lot. We learned a lot about = accounting.=20 We learned a lot about manufacturing. And we grew consistently. The = opportunity=20 to do everything is something exciting in itself and very satisfying, an = important part of starting a business. That part I would recommend to = anybody=20 [who] had the opportunity to do.=20

    DKA: You said you learned a lot of things. I assume the things you = learned=20 helped you succeed.=20

    KO: Well, we learned a lot of things. Some were useless and some = were... how=20 to keep pigeons out of the building. Not particularly useful afterward. = But just=20 understanding how accounting systems work and personnel problems. How = you hire=20 and how you fire. How you purchase. How other companies work. [These] = are all=20 things you learn when you're small.=20

    DKA: What did your customers think?=20

    KO: The potential customers were readily easy to define because they = were=20 people who wanted high speed circuits. And so we went to trade shows and = we'd=20 call on the people we knew. It was always touch and go. We were = profitable every=20 year but you very rarely have so many orders that you feel completely = secure.=20 It's just the nature of business.=20

    DKA: What was it you were offering?=20

    KO: To start with, we were offering modules for laboratory use that = were=20 faster than anyone else [was making]. So people would buy our modules = and=20 experiment with high speed computer technology or test devices they were = building. This was our first offering. Later on we offered modules = people would=20 use to build things they were going to sell to others. Then we offered = computers=20 made out of these modules; called the first one our PDP-1.=20

    DKA: What else do you want to say about these early days? How did = your wife,=20 for example, feel about your going into business? Did she feel worried = that you=20 were taking this big step? Or was she supportive?=20

    KO: I can't tell you why, but she was never worried. We probably came = from a=20 different world. You never had much in the Depression and you didn't = worry about=20 having much. She had come through two wars in Finland, and much of her = life,=20 never had enough food to be completely satisfied. You never really = worried about=20 failure because you know, it didn't make any difference. So we didn't = have=20 anything, we didn't worry about anything. So that was part of it. The = pay was=20 what I had gotten at Lincoln Laboratory. So the risk we never really = worried=20 about. Now the risk in business is the different risk. And that is, when = companies fail, it's a miserable death. They fail and if anyone has an = emotional=20 involvement, it really is agony. But it's not the risk you think of = normally,=20 the financial risk or something. It's the risk of watching something = die. So=20 that risk we didn't worry about. We didn't think of.=20

    DKA: What was your own personal challenge? What was the demand on you = that=20 was most acute in those very early years?=20

    KO: The excitement, the fun, the thrill was to do everything. This = also, of=20 course, put certain demands on. There's only so many days of the week = and so=20 many hours in the day. Balancing that with family was always a = challenge. But it=20 never really got out of hand. I never really felt overworked. In general = it was=20 very exciting. I came home for supper every night, spent the weekends = with my=20 kids. There's always a list of things to do. And you just systematically = go=20 through the list. Early in the morning, walking in the woods. Just = generating=20 that list of things, keeping on balance. And then click them off one at = a time.=20 And as long as it's approached systematically there's no great tension. = If they=20 ever jumble up in your head and you get behind there can be quite a bit = of=20 tension. But in general it's quite clear what should be done. And not = trying to=20 do things that are impossible. And not worry about the pressures that = other=20 people would like to impose on you. The pressures on someone in business = are to=20 take part in every outside activity, go to dinner for something or other = every=20 night of the week. Most of them are useless and most of them have = nothing to do=20 with the business. Saying no to them is a major operation, a major key = to=20 survival. Another story which I tell when people ask me what they should = do to=20 learn to run a business, I say, jump at the chance to run something. If = you're=20 working, offer to run the cafeteria, the parking lot, things nobody else = wants=20 to run. If you have an opportunity to run something in town or in = church, just=20 run something. Manage it. You learn to manage by managing. But don't = think=20 management comes from a book and then you're suddenly going to do it. I = told the=20 story of how I got started from an MIT point of view. The other story I = tell was=20 that when I was 30 I was drafted to run the Sunday School of a large = Boston=20 church which to me looked kind of large and stuffy. Everybody was old. = They must=20 all have been fifty. Some of them quite a bit older. I accepted the job. = The=20 first thing I did was go to the Lexington library and take out all the = books on=20 management. I can remember what I learned then. I can't remember = anything I=20 learned since then. But approaching every job because it was a = management job,=20 and learning, it's an excuse to learn something. My taking that job was = probably=20 a key part in learning and being interested in taking on a management = job. The=20 other thing is that getting people to work enthusiastically is always a=20 challenge. You obviously can't do very much yourself. But if you can get = others=20 to feel it's their job, their invention, their contribution, they can = get an=20 enormous amount done. Making sure they feel that way is a key part of = it.=20

    DKA: I wondered if you wanted to talk this morning about the team of = people=20 that you brought together to work with you and the principles that you = had them=20 follow as you moved towards this first product, the PDP-1.=20

    KO: Developing and managing an organization like Digital is a = compromise or a=20 set of paradoxes, or conflicts between leadership and giving = responsibility to=20 others. It's obvious that leader, myself in this case, can never be = expert in=20 everything. We have to be dependent on those people who are. They = obviously have=20 to do the design, set the goals. They have to have the motivation that = comes=20 from them, setting the goals. And yet we have to have a common goal and = that's=20 obviously the job of the leader. What I did for good, or for not so = good, is=20 probably demonstrated many of our products. Here we have our PDP-1. The=20 background I had, the experience I had was the design of circuits, the = design of=20 logic, how you did arithmetic with transistors. But early in the history = of=20 Digital we could hire people who were expert in that. The area we = couldn't hire=20 people were the making of power supplies, the putting together the = packing, the=20 industrial design. So, in that case, I gave the responsibility for the = things I=20 had been expert in to those who[m] we could hire.=20

    DKA: Ken, what would you say was the overall goal in making the = PDP-1?=20

    TO=20 CONTENTS

  • Digital's PDP-1 Computer

    KO: The goal of the PDP-1 was to introduce a new type computer to the = world.=20 In the tradition that was developed at MIT where the computer was very = simple,=20 very fast, relatively inexpensive. In this case, [the price was] = $110,000 with=20 only 4,000 words of memory. Because it was simple, easy to use, = interactive with=20 the cathode ray tube and light pen, it could be used by an individual. = Someone=20 could afford to sit there and use the computer like we do [with] a = personal=20 computer today. You could also use the same equipment interactively with = equipment, that is, with a machine or a telephone system because the = price was=20 relatively low. $110,000 thirty years ago was a lot of money, but = computers then=20 cost one, two and three million dollars, so it was relatively = inexpensive. And=20 it did open up new applications that people hadn't thought of before.=20

    DKA: Why hadn't anybody done this?=20

    TO=20 CONTENTS

  • The Concept of an Interactive Computer

    KO: First of all, we had experience with the technology. After years = at MIT=20 it was just natural to us. The concept of an interactive computer was = strange.=20 Some people thought it was wrong. Almost spoke in ethical terms. = Computers are=20 serious, you shouldn't treat them lightly. You shouldn't have fun with = them.=20 They shouldn't be exciting. They should be formal and distant with red = tape=20 involved. That was the atmosphere at the time. So it was a strange idea. = The=20 other motivation we had was that we believed computers should be fun. = They were=20 exciting. They could do so many things. The opportunities were just = without=20 bounds. This was a great motivation in building a computer. But it was = not=20 commonly shared in the industry. Now, the other reason of course, was = that using=20 vacuum tubes in the older technology, the machines were big. They were = huge. And=20 they were expensive.=20

    DKA: As you look at the history of computers in this period, you see = people=20 focusing on operating systems and getting batch processing and more = efficient=20 use of computers. Was there a thought that this wouldn't be an efficient = use?=20 Was that part of the concern?=20

    TO=20 CONTENTS

  • Real-Time Data Processing

    KO: The original computing was based on the way people had done = computations=20 before. You'd collect all the data, bring it together, process it and = send the=20 answers back. The idea of processing it, real time, took a long time to = develop.=20 In the world of commercial processing, it's just in the last few years = that=20 batch processing has started to disappear. The replacement for it is now = called=20 transactional processing, where if you make a transaction with a bank, = it is=20 instantly taken care of. Your accounts are updated. And you could have a = new=20 transaction almost immediately. So, 30 years later, it has influenced = the=20 commercial market. That was dependent on software and large computers = that were=20 fast.=20

    DKA: Did this PDP-1 conceptually have any relationship to what you've = just=20 described?=20

    KO: The PDP-1, we like to think, along with the circuits we were = selling, was=20 a vehicle for introducing the MIT ideas into the rest of the world of = computing.=20 The idea of fast simple machines was strange in those early days, and we = like to=20 think that we helped change the world. The MIT tradition came about = because we=20 were working with physical phenomena. In aerodefense we divided the = country up=20 into 16 bits and that was close enough for defending it. For most = physical=20 phenomena, 16 bits is enough. The rest of the world ridiculed anything = [with]=20 that short word length. And, as you well know, 16 bits is the most = common of the=20 powerful personal computers today, and up until just lately, most = minicomputers.=20 So even the word length probably came out of the MIT tradition, and the = PDP-1=20 tradition.=20

    DKA: Ken, did your whole team have a common feeling about this = objective, or=20 was there difference in opinion as to which direction you should go in=20 computers?=20

    KO: When it came to details, we had all kinds of opinions. We spent a = lot of=20 time arguing, and openly exposing all the differences. But the firm = belief that=20 the world needed fast, inexpensive computers that were interactive was = just=20 accepted, without question. We also believed, now maybe this sounds a = little=20 naive, that every computer had to have a cathode ray tube. I mean it was = inconceivable. But that you wouldn't have one with a light pen. Also, = every one=20 had an audio output. I think somewhere, under this console there's one. = Because=20 it goes with interactive computing. If you lost interest in everything = else and=20 you were alone with the computer, you could always write music at night. = So that=20 sort of thing was just accepted in the environment. Along with the idea = that=20 computers are fun, exciting, and that anybody can learn them. From a = young child=20 on up. When we asked for money we may not have mentioned that computers = should=20 be fun. But every time we reported to them, we reported that computers = should be=20 fun and exciting and therefore very productive. The demonstrations we = put on in=20 the stockholder's meeting of American Research, which owned almost all = the=20 stock, demonstrated the fun and the interactiv[ity] and the = responsive[ness] and=20 the productivity that came from it. So we openly presented this.=20

    TO=20 CONTENTS

  • M.I.T. Students Learning Computing on Digital's = PDP-1

    KO: When we were at MIT, we at Lincoln Laboratory gave to the = educational=20 part of MIT at the TX-0 computer. Then when we started Digital we gave = one of=20 our first PDP-1's to MIT for use by the students. The students could = then have=20 opportunity to use that machine anytime, 24 hours a day, and they could = sign up=20 for it months ahead and do anything they wanted. They learned more about = the=20 computer and how to do things with it than probably anybody had before = that=20 because you had dozens of bright people spending all hours of the day = studying=20 this. Out of this came what we know of today as video games. We had = played fixed=20 games before in an oscilloscope like kalah and the Asian games that you = could=20 demonstrate here. But what we know of as video games came out of that = group at=20 MIT. These always made great demonstrations. Spacewar! being the most = exciting=20 one. It's] still showing in the [Computer Museum] museum. And very = close, you=20 can just see that out of that came what we know today as video games. = This made=20 a great demonstration for stockholders. We also played music. I think we = had a=20 four part Bach we could play. We did feel passionately that we had an = approach.=20 We were too much scientists to say we invented anything or it belonged = to us, or=20 uniquely ours. But we did have this missionary zeal to introduce these=20 technologies to the world. And we got great satisfaction to see it = develop, both=20 in the world of human interaction and machine interaction. Interacting = with=20 machine tools or laboratory devices or telephone lines. It was the goal = we set=20 out to do. It was a goal we formally stated. And this worked out very = well.=20

    DKA: Ken, you talked a lot yesterday about the modules that DEC first = made as=20 being your initial product. What's the relationship between the modules = and this=20 initial computer?=20

    KO: Digital devices are made up basically of just a small number of = circuits.=20 Flip flop holds two states which it holds the ________. Gates, you'll = put=20 various combinations of signals in and if the right combination is met, = the=20 signal gets through. Then there are amplifiers and things to allow you = to run by=20 drive things, or run distances. In making up modules out of these = building=20 blocks our original customers could build up computer like devices, or = actually=20 computers. We then used them to build the PDP-1. This is the back of the = panel=20 but those modules plugged into the back. [POINTING OUT PARTS OF THE = MACHINE]=20 These were all wired up and soldered by hand. But the correct modules = plugged in=20 the back. If they were defective you could change them easily. If you = designed=20 them once and tested them thoroughly and then you used the same ones = many=20 places. I think there were probably 20 across, 20 times 20 or 400 = modules. Maybe=20 it was 600.=20

    DKA: How did you make it so fast?=20

    KO: The approach to the circuits which evolved at MIT was to set = about to=20 make the circuits fast and then the computer simple. It meant very = expensive=20 circuits. The transistors cost $12 each. Maybe by the time of the PDP-1 = they=20 were down to $6 or $8. That's very expensive. So that the circuits in = the=20 modules were very expensive. But the result was basically a very simple = machine,=20 and one we could readily build and that operated very fast.=20

    DKA: Tell me about the sales of this machine. Who wanted it and what = did they=20 think about it when they got it?=20

    TO=20 CONTENTS

  • Digital Customers Developing Software

    KO: We had contracted with some professors at MIT to do the software = for us.=20 They didn't come through. We then announced the machine as the machine = without=20 much software. So, this was a challenge to a group of customers who = wanted to do=20 their own software. The applications normally went to the scientists. = [REFERENCE=20 TO MACHINE ON EXHIBIT - This machine here is not as good looking as the = original=20 one. The frame over the cathode ray tube is for hanging a camera on. = This ugly=20 bracket was for holding a camera. And it probably was used for high = energy=20 physics experiments. They could take the data, project it on a tube and=20 photograph it. And that was the general application.] The American = International=20 Telephone & Telegraph bought a large number of them to collect = teletype=20 messages, stored them and then distributed them to the phone lines that = were=20 free at an optimum rate rather than storing on paper tape and doing it = manually=20 like they did before. So that was an obvious application for it.=20

    KO: [Our early customers] jumped at the chance to [develop their own=20 software] for their special projets.=20

    DKA: One of the things that DEC is well known for is working closely = with=20 customers. Was that part of what you had to do then almost as a = necessity?=20

    KO: One of the exciting things about this business is that you have = such an=20 interesting, diverse set of customers with very interesting problems. = Each one=20 is different. Most of them are quite exciting, all the way from physics = to=20 telephone controls...education to medicine. Working with them is = important.=20 Through our history, many of the ideas came from our customers. We = obviously=20 were never limited by my knowledge or my ability to come up with ideas. = But we=20 also were not limited by even the people who were responsible [for] = various=20 parts of the technology. Customers, when you're close to them, often = have ideas.=20 Sometimes [they] actually do the development, and are excited if we take = them=20 and produce them. Being close to the customer was very important.=20

    DKA: Is that a sort of philosophical principle that runs through your = company?=20

    KO: It's much simpler than that. It's an obvious philosophy. It's = very=20 productive. Computer people sometimes think more highly of themselves. = So=20 there's a tendency to not be humble enough to take someone else's idea. = But the=20 idea itself is so obvious you should just jump at any opportunity to get = better=20 use out of your computers.=20

    DKA: You were mentioning several of the uses of the PDP-1 and you = mentioned=20 the telephone usage and the scientific usage. What else were people = doing with=20 these new machines?=20

    KO: Some of the things we didn't know. Some of the things people = wanted to=20 keep to themselves, either because they were government applications or = things=20 they wanted to keep private. In the laboratory area you could do things = that=20 were so tedious before. One of the first applications for this type of = computer=20 that gave dramatic results were to automate the x-ray defractometers. = These are=20 devices where people would put a specimen in, take pictures and then = take days=20 or weeks analyzing the results. The computer would give them results=20 immediately. If the experiment wasn't done right, they could do it over = again.=20 But they could run through many experiments very quickly just because = the data=20 was processed automatically. Another time we put these machines, or one = like=20 this, on a Coast Guard ship doing oceanography work in the ocean. It=20 revolutionized their work. They could collect data, process it, [and] if = the=20 data wasn't coming out right, they could do the collection over again. = The=20 thought that devastated them all throughout a cruise was, if things = weren't=20 going well, they wouldn't find out until they were home and it was too = late to=20 redo the data. So, it just revolutionized their enthusiasm. [There were] = just an=20 infinite number of applications.=20

    TO=20 CONTENTS

  • Selling Computers to Government

    DKA: We hear so much about the Defense Department in the early days = of=20 computers, particularly the large computers as being their principle = sponsor.=20 Did this new, less expensive device, open up new areas of markets for = you or was=20 it still mostly defense and defense-related business that you got into?=20

    KO: When we started, we had the policy that we wouldn't sell to the = Defense=20 Department. [For reasons of] the accounting they demand, not [that we = were]=20 pacifists. And the products they develop are just contrary to commercial = activities, the commercial way of doing business. We felt and still = feel, still=20 very clearly, that doing business with them hurts one's position in the=20 commercial market. Now we do a lot with them, we sell to them freely = because=20 it's our duty to do so. We normally do it on commercial terms, because = we still=20 are afraid of changing the nature of a commercial company if you aim = your=20 business to satisfy their way of doing business.=20

    DKA: Can you characterize a little bit more the difference between = working=20 for a defense market and a commercial market? Is it flexibility and = freedom or=20 is it something else?=20

    KO: The government business...the way Congress sets it up and the way = that=20 people want to run things...worr[ies] about you making profit. They = worry less=20 about what they get. That's absolutely contrary to the commercial = activity. Most=20 commercial customers want to do business with you only if you make a = profit, and=20 like you more if you make a good profit because it shows you're doing = things=20 wisely and you'll be around for awhile. The commercial companies always = want to=20 buy the best product and they have learned, those that are successful, = to trust=20 their suppliers. There's a relationship between the suppliers. The way=20 government business is set up it's always distrust. The supplier is = always the=20 enemy. There's always someone who wants to find something. You're = treated like a=20 criminal. It's just not a good way to do business. Their accounting is = contrary=20 to common sense and contrary to commercial business practice. If you set = it up=20 that way you're really not competitive in a commercial market. So we = originally=20 made the rule we would sell nothing to them. They, in time of course, = insisted=20 because they needed what we had. And we, to this day, limit our business = with=20 them to commercial terms.=20

    DKA: In the early days of computing when the business was government, = [that=20 must have been a] risky principle, wasn't it?=20

    KO: No, it was just common sense. It was risky to do anything else. = The=20 financial community didn't understand this. We said we wouldn't do = business with=20 them. The financial community thought that if you took government money = to do=20 your research it was free research and then you had a head start. It = doesn't=20 really work that way.=20

    DKA: Why not?=20

    KO: The red tape they put in and their goals are just contrary to = commercial=20 goals. Therefore they're always behind.=20

    DKA: As I look at this equipment, I'm curious to know how this = differed from=20 what other companies had out. Was this innovative? Was this bringing = something=20 to market the market wanted but couldn't have before?=20

    TO=20 CONTENTS

  • Unique Features of the PDP-1

    KO: The thing that's unique about the equipment is the fact that it's = simple,=20 very fast, and interactive. The market wasn't demanding this. People had = never=20 seen it, didn't know about it, and didn't ask for it, with very few = exceptions.=20 Some of the students who had used it at MIT when they got to their new = company=20 they said, where's the PDP? Our generosity at MIT paid off very well. = But in=20 general, the market wasn't asking for it. Market surveys came to the = conclusion=20 that people wanted exactly what they had. That's because that's all = they'd ever=20 seen. And that's why you get into trouble if you believe market surveys = all the=20 time because they never come out with anything useful on new products. = So we had=20 to, in general, sell the idea to most of the customers. We are not = considered a=20 marketing company because we don't spend money on _______ markets. We = don't=20 exaggerate. We try tediously, sometimes boringly, not to mislead or be=20 dishonest. Therefore we're not considered marketers. We refuse, when we = have the=20 opportunity, to sell to someone who doesn't need our equipment. If it = won't do=20 the job for them, or won't do it optimally, we don't want to sell it to = them.=20 That's considered non-marketing. But on the other hand, the goal we set = about=20 was to sell this product to people who needed it. It often didn't look = flashy=20 but it worked for 30 years. Now there was a certain other flashy = marketing which=20 did help. It turns out that physicists love color and a little bit of = spirit.=20 Now this machine [THE PDP-1] today is a little naive from industrial = design. But=20 the color, the consistency did give a little more pizzazz and jazz than = most=20 computers; people loved it just for that sake. So that was part of the=20 marketing. The shape of the monitor in that time was unusual. The = console really=20 was designed to look good, to have something you like to work with and = would=20 attract people. That was part of the marketing.=20

    DKA: Were you involved in the looks of this stuff yourself?=20

    KO: My job was to make sure we had goals and to make sure that people = knew=20 them and that we were all running in the same direction, and running = with=20 enthusiasm. And then quietly to pick up the pieces that other people = left=20 behind, or that we didn't have people to survive. The mechanical design = was one=20 of those. So, yes, I was involved in that. It normally is not one of the = things=20 people notice, but because I did some of it, I of course will talk about = it.=20 It's relatively unimportant. But some of the ideas are interesting. = Early in=20 computers people built a console on the table with all the instruments = and=20 lights in front of them, very much like you see in a power plant. Our = first=20 PDP-1 was that way. It had all the controls on the table, and big heavy = cables=20 going over the computer to connect the tube. Finally it dawned on me. = Maybe the=20 first in the industry. Something very simple. In the control, in the = power=20 plant, they have a console that way so they can look over at that big = array of=20 lights and heaters(?). Sitting at a computer console, the only thing you = look=20 over at are dull racks. So the whole thing's stupid. We got the idea of = putting=20 it on the end of the computer. After that everybody followed. Now that's = a=20 trivial thing, but maybe that was a contribution I made. You never can = say for=20 sure because somebody else might have done it first. We got the idea = from them.=20 But that's a small contribution. Picking a color, picking the logo type, = this=20 sort of thing. There was no one else to do it so those were left for me. = I did=20 have a little machine shop practice so I knew how to design. I know how = to use a=20 drafting table.=20

    DKA: What's PDP mean? Where did that come from?=20

    TO=20 CONTENTS

  • PDP - The Origins of the Name

    KO: When we were almost finished with the computer, before we had = picked a=20 name for it, we had a request from the government to build a machine to = look for=20 earthquakes. To collect the seismographic information. We didn't quite = believe=20 the story that people were that interested in earthquakes but we were = willing to=20 let it go at that. The Congress, in their wisdom, said that no more = computers=20 will be bought until all the computers in Washington are used 100 = percent of the=20 time. Now it seemed unlikely that earthquakes would wait until the = accounting=20 machines were unused before they went or the Russians would be so = helpful as to=20 not try out a bomb at the time when the machines were available. So they = needed=20 a machine but they couldn't buy a computer to run their seismographic = machines=20 just because Congress had this rule. We said you could call it something = else.=20 It's really a Programmed Data Processor. So we called our machines, PDP, = for=20 Programmed Data Processor. They were able to buy the machine, hook up = their=20 seismographic devices, and not [have them] break, and not be challenged = on their=20 purchasing of the computer. It was a good name and that's where PDP came = from.=20 The job of the leader is to fill in those tasks which no one else is = going to=20 do. But never claim credit for them because that's just contrary = leadership.=20 You're supposed to be getting everybody else to work and if you pick up = some of=20 the pieces, you should never brag about it. The story I tell people who = get=20 confused on this is [a] fable. It used to be in THE FIRST GRADE READER, = "The=20 Turtle That Wanted To Fly." He talked the crows into putting a stick = between=20 their mouths, and he held on the center of the stick and flew with them. = Someone=20 on the ground said, "That's a clever idea, who thought of it?" He = couldn't keep=20 his mouth shut. He had to say, "It was me." So my advice to people who = want to=20 be leaders is, remember the task as a leader is not to claim credit, but = to be=20 the leader and get the job done. The other advice is you've got to make = sure=20 everything gets done. And so the logo type, we... I did. I did it myself = and I=20 can tell you pretty much where it came from. The first one had a = vertical of=20 DEC. It was very clever.=20

    TO=20 CONTENTS

  • DEC's First Trademark

    KO: Our first trademark was the vertical DEC. I got the idea for that = right=20 from the cover of a magazine. It was a magazine that's probably changed = its name=20 two or three times since then. I don't remember which one it was. But = having the=20 "D" reverse color at the top of the square I thought was clever. So we = simply=20 adapted it to this. We silkscreened our printed circuit boards. We made = up=20 screens with a trademark on it and the name on it, and then we did it=20 everywhere. We put it on our doors. We put it on our used Volkswagen = bus. Once=20 the screen was going, you know, we'd put it up everywhere. In time we = found that=20 people called us Digital and not DEC. So we made the decision quite = formally to=20 change our name to our shortened name, our trademark to Digital instead = of DEC.=20 Another magazine, I think it was called machine design at the time, had = their=20 name in blocks like that. It looked attractive. Those days lower-case = letters=20 were the smart thing to do. So we designed this logotype, and it became = the=20 standard. It might also have been that the rest of the machine design = magazine=20 used the rest of the word like this. But the ideas came right from the = cover of=20 a magazine.=20

    DKA: I've often heard that it had something to do with the storage = register=20 of...=20

    KO: No. Not at all, not at all. Now one time our experts suggested = that lower=20 case letters were passe. That these were hard to read. There was a newer = alphabet. They proposed a variation of this with more modern letters. We = brought=20 it to our Board of Directors and I said we want to do it and they said = yes. That=20 night we happened to have dinner with the wives and the directors. One = of the=20 directors said, "I'd like to call the meeting to order and re-open this=20 question." He was sitting with my wife...my wife doesn't take part in = business=20 except a few times and [this] being one of [those times], where they = proposed=20 that that may not be the best, but it's Digital, and they proposed we = stay with=20 the old one. So we stayed with Digital.=20

    DKA: What did your wife have to do with that?=20

    KO: At dinner, the question of changing it to a more modern type face = came up=20 and she and this director connived together to raise the issue again. = When they=20 called for a vote, wives don't have any vote, but we counted 18 together = in this=20 informal vote. I said I claim 19 votes and we're going to stick with the = original decision. The next day of course we formally went back to the = original=20 and never changed it since.=20

    DKA: On signs, there was another story. We put signs outside of = buildings and=20 after awhile the sign people had very complicated signs. They had = Digital=20 Equipment Corporation, and then the name of the location and something = else on=20 it. It didn't feel right. So, as I drove with my wife we looked at = signs. We=20 concluded that the most effective signs like Mobil or Esso were just the = simple=20 name. So we decided to use that, and that alone, for our signs. It's = very=20 presumptuous because not everybody knew who Digital was. But challenging = them by=20 being a little presumptuous was also a fascinating idea. It is sure more = effective to just [use] that simple word than have a whole lot of other = words.=20 So the other story is people still call us both DEC and Digital. There's = always=20 the challenge, the question by our people from all over the world, we = should go=20 back to DEC because their people use that. I of course answer, no way. = We've got=20 so much invested here I'm not going to raise the issue again. If you = look=20 carefully, it's about 50/50. And if you listen to me, I may say half the = time=20 "Digital" and half the time "DEC." It's not worth changing. The press = uses=20 whatever they want to use regardless of what we call ourselves; they're = going to=20 use what they want anyway. So, that's it for a long time to come. If = people say=20 they propose changing it, I'll say you can bring it to the Directors. = You won't=20 win.=20

    DKA: So those little rectangles behind each lowercase letter do not = have=20 anything to do with modularity?=20

    KO: No, no they just look good from an artistic point of view.=20

    DKA: Ken, let's go back and talk about the software issue for the = PDP-1. You=20 mentioned initially it came out with very little software. As the = machine=20 started to sell, what was the company's position on software = development? How=20 did you tackle that problem?=20

    TO=20 CONTENTS

  • Software for the PDP-1

    KO: In time we invested as much in software as we did with hardware. = For a=20 long time we had the same number of people doing both software and = hardware. The=20 concentration on the software was never in applications but on the = systems to=20 make it easy to write software, also on the discipline to make the = software very=20 robust and very reliable. One of the results of this is that almost = every year=20 of our history, except the very start, there has been someone who's had = a faster=20 computer at a lower price than we have. By concentrating on the = discipline, both=20 the hardware and the software, but even more on the software. Extreme=20 discipline. And every year making it better and better. Keeping the = hardware=20 architecture roughly the same so that things were compatible. And just = building=20 on more discipline and more security and more robustness. In time every = one of=20 those quick upstarts who offered something faster have disappeared. The = belief=20 there is that in computers you really want, above all, reliability. You = don't=20 want to lose things in a thunderstorm. If everything goes the power = disappears.=20 You want all the things you worked on, all your data tucked away nicely = so when=20 power comes on you can get it back again. This has been our approach in = software=20 and the thing that's made us unique in the last number of years is this=20 continuation in an organized way of discipline documentation and = sticking with=20 standards. So our approach to software has been quite consistent, but = it's been=20 in the area of discipline and documentation. For many years we made the = same two=20 computers, the PDP-8 and the PDP-11. We kept that design consistently so = that=20 software the customers wrote would continue to work on newer models and = the=20 software we wrote would continue to work and get more and more robust. = By that=20 we mean, regardless of what happens it will be safe and secure. This was = dependent on a high level of discipline and organization and = documentation.=20 Software is something which you can't look at and understand right away = because=20 it's a series of numbers, and pages and pages of them. You never know if = parts=20 are even used. You might have a million instructions and in no way can = someone=20 tell whether parts of it or most of it are just left over and never used = at all.=20 When people have trouble transferring software to new machines, they = hate to=20 admit that it's often the lack of discipline in the software. So by = adding this=20 discipline and the systems to it, we've gotten people to be dependent on = our=20 machines and because they have the commitment from us that we will stick = with it=20 and continue to maintain it and with discipline. When the Russians want = to steal=20 a computer they want to steal a VAX. The reason they want to steal a VAX = is=20 that's where the software is. The reason the software is there is that = it's=20 disciplined and easy to use, and built upon for many years. In today's = world of=20 software, we have our VAX system. We call the software VMS. We have = basically=20 one system. It has about a 12-foot long shelf of books to describe it. = Anything=20 you want to do you'll find in there if you look long enough. Every = change is=20 tested and retested, and everything is safe and as secure as we can do. = The=20 problem with this is that it is a 12 foot long shelf. To take advantage = of it=20 all takes some time. Another system was developed even before we = developed VMS=20 called UNIX. It was designed in reaction to complex business systems. It = was=20 designed for one person to work alone. It was designed to be without = discipline,=20 freeflowing, and it became very popular in the academic world because it = was=20 easy to learn. It's easy to do simple things in workstations. And we've = sold=20 more of that than anyone else has partly because it was designed for our = machines. We've been the biggest player in that market, and encouraged = it=20 because there's a place for that casual undisciplined work, and a place = for the=20 discipline work. So today, UNIX is very popular for workstations where = one man=20 works alone, where if he loses everything because of a thunderstorm, = he'll feel=20 bad. He's lost a whole day's work. But it's not the same as a bank who = just=20 under no circumstance can ever, inconceivably lose anything. These are = the two=20 different realms of discipline and security. Now the promise is we = promise that=20 we'll make UNIX secure, and robust, and formal and disciplined, with a = lament=20 secretly that when we do, we will lose that free-wheeling, fun, casual,=20 easy-to-use UNIX system that made it so popular.=20

    DKA: I'd like to take you back to the PDP-1. One of the things that = you've=20 said is you learned a lot from your customers. What went right with the = PDP-1,=20 and what went wrong? What did you want to do different as you moved? Or = did you=20 just want to do more of the same? As you saw the company growing, which=20 direction did you want to take it?=20

    TO=20 CONTENTS

  • Managing the Success of the PDP-1

    KO: We had large orders from ITT, for telephone, telegraph type = systems. They=20 decided to go out of that business, which left us with more production = than we=20 had orders for. At the same time, NASA was getting ready for their big = projects=20 and they wanted to buy 100 computers. And I said, no. The engineer who = designed=20 it left us after that because 100 machines would have made it a great = success.=20 But growing that fast would have ruined the company. We were making 2 or = 3 a=20 month. Getting an order for 100, we couldn't have tolerated. Then if = that order=20 disappeared, we'd be in terrible shape. So I said, no, we won't bid on = it. It=20 must have taken years before NASA got over that.=20

    DKA: That must have been a hard decision, Ken?=20

    KO: No, it was an easy decision. Suicide is something you don't have = to do.=20 My first rule of business [is] you don't have to commit suicide. People = think=20 it's a hard decision. [It's] not a hard decision at all. It's like = hanging over=20 a cliff on a narrow piece of rope. That's not a hard decision. Some = people felt=20 it was hard because the short term result could have been exciting. But = we had=20 to grow slowly and consistently or we would lose everything. Now each = step of=20 improvement afterward was quite obvious. The technology got better = consistently=20 as time went on and we understood more how to build machines.=20

    DKA: What was getting better? Can you be a little more specific?=20

    KO: The old typewriters which we bought were crude and expensive. As = we got=20 better typewriters it just made the computer available to the masses. = The=20 transistors got better. Eventually, circuits came into play. We learned = how to=20 put more on a printed circuit board. We learned better ways of packaging = and as=20 computers developed, more and more of every kind of component with = better=20 quality and lower price. That meant we could cut the cost significantly = and=20 increase the power. The industrial design was always important because = people=20 did buy [computers] for looks. They liked something, it was easy to use, = it was=20 very productive, but the looks were important. We always had either a = rented=20 industrial design[er] or part of our own staff. And this [THE PDP-1] we = thought=20 was quite good looking. The console had the same general shape as the = display.=20 And [the] layout of switches was studied very precisely. We made special = knobs=20 for it. It was readily accessible to the tape reader. The tape storage = was=20 obviously patched in later on. But it did make a very efficient way of = doing it.=20 The punch tape came out here. It made a very efficient and attractive = system, we=20 thought.=20

    DKA: Ken, was it all paper tape or did you have magnetic tape at all = for the=20 PDP-1?=20

    KO: The software people brought to the machine with them was paper = tape.=20 There was magnetic tape for storage but people didn't normally bring the = [magnetic] tape with them. They usually brought the paper tape, or = they'd bring=20 a tray like this [DEMONSTRATING PAPER TAPE TRAY]=20

    TO=20 CONTENTS

  • Running a Program on a PDP-1

    DKA: Now what would it be like for somebody to run a program? How = would they=20 set it up?=20

    KO: [DEMONSTRATING USE OF PAPER TAPE READER] They would take the tape = out and=20 put it inside here, string it through the tape reader and set it up so = it would=20 fold here and the tape would come off and this fanfold. When he left he = would=20 have all his tapes put in a tray like this or several trays like this. = Any new=20 tape he generated while working on it, he put in trays also.=20

    DKA: What was he doing on the control panel below? What would a = programmer=20 do?=20

    KO: As he was debugging his system, he could watch what was happening = all=20 through the computer by looking at the lights. He could introduce = information=20 through six switches which the computer would sense. If you wanted to = play a=20 game, you'd play it by entering information in the switches. The = computer would=20 light up lights here, special program lights to tell him various things, = or tell=20 him what was going on if he was playing a game. These were just start = and stop=20 and continue. [POINTING OUT SWITCHES] If he examined, he'd press this = one that=20 would stop the program. He could see from the lights what was going on. = He could=20 step the program along one step at a time, and just watch what was = happening=20 with his program as he went through. In time, slowly over a period of = many=20 years, we got rid of all of this. Now people don't need that at all.=20

    DKA: What was the relationship between what people saw on the control = panel=20 and what they saw on the CRT?=20

    KO: The control panel showed exactly what was going on in the = computer. The=20 CRT would present pictures that someone worked hard to generate. At that = time=20 you couldn't debug from what you saw in the cathode ray tube. Today you = don't=20 need this because the cathode ray tube does it so well. We always had = very good=20 documentation. [For the] PDP-1, we worked very hard on a manual, and it = wasn't=20 very big. People loved it. The competition thought they'd do the same = and were=20 dismayed to find out how hard it was to write a simple small manual.=20 Unfortunately, not being a normal user, there's always one thing we left = out. We=20 never told which way you put the tape in. I never figured it out. I = never=20 remembered it. So there's one weakness in our literature. There's = another=20 literature story: by policy, we would write everything we do in book = form for=20 the customer. We formally decided we would print things in a paperback = format,=20 and put in everything we knew about a computer. We put a manual out that = told=20 absolutely everything. At that time we could have that book printed for = twelve=20 cents. Our experts in publishing thought it was a disgrace. Paperback = books were=20 inexpensive paper, obviously not glossy and polished photography. But = the=20 difference was, a brochure which could say almost nothing would cost a = dollar. A=20 paperback would cost twelve cents and we could tell everything in it. = One time=20 after meeting on a Friday afternoon, my brother, [STAN OLSEN] to = demonstrate=20 something, ripped one of the paperback books apart, and laid it on a = table.=20 [Then] he ripped some literature apart and laid it on the table. Come = Monday,=20 the janitor had thrown the ripped literature away but carefully put the=20 paperback together. Everywhere you went in the world -- India, China --=20 engineers had those paperbacks on their bookcase because they were that = popular.=20 We gave them away freely and we got a lot of information out. They were = very=20 valuable even to those people who couldn't afford a computer; [they] = educated=20 them about every single detail of our machines. So, since then we've = been free=20 in literature; we print everything we know [about the system] and [the=20 handbooks] make a real contribution to the customer.=20

    DKA: Is DEC more open with its information about machine design than = other=20 companies?=20

    KO: Probably so, because of [our] academic background. But even more=20 important, it takes a lot of discipline to get people to write. After = you're=20 finished with a job, it's very hard to write about you've done because = you're=20 ready to go on to something new. Getting people to write down all they = know=20 about the project they worked on takes discipline and effort. It's one = of the=20 things we're never satisfied with but we keep trying to do a better job. = As=20 [computer] software became developed, it [software] helped him [the = programmer]=20 find the problems and correct them in this interactive way. This is very = much=20 more productive than getting your results back once a day. Getting your = results=20 immediately allows you to fix it immediately. As you get clever, it = allows you=20 to use the computer to help you fix it.=20

    DKA: Ken, did you use card readers at all with the PDP-1?=20

    KO: In general, we used paper tape. The scientific users used paper = tape. We=20 always had a card reader available because commercial users or some = people whose=20 data came in punched card form did have to have access to punch cards. = But the=20 paper tape was much easier to use for the casual user.=20

    DKA: [Was] memory an important consideration of the PDP-1?=20

    KO: Yes. The memory, for many years, was the limiting factor in a = computer.=20 We decided to build this PDP-1 a little earlier than we had budgeted = because our=20 friends at RCA called us one day and said we have a thousand word memory = the=20 customer ordered and doesn't want. We'll sell it to you cheap. So we = quickly=20 bought it and said, now we've got to make our computer. The computer was = designed around the thousand word memory. That was the impetus to get = going. We=20 sold that 1000 word memory many times. It deserves a place in the = museum. We=20 don't know where it is though because people would buy the machine with = a 1000=20 words. Before they got it delivered they needed more. And they needed = eight;=20 four, eight or twelve thousand words. And we'd then sell the machine, = that 1000=20 words over again to someone else. Now those numbers are ridiculously = small. But=20 the price and the size of the memory has been growing, the size has been = growing, the price has been going down dramatically. It probably is the = most=20 serious limitation in the computer at any one period in the history. = Every=20 component has to work and be available. The typewriter, transistors, and = the=20 memory. Every single piece has to be available. Probably the hardest = one, the=20 most serious limitation in any time was the memory. So as we learned to = do=20 memories better, as the people that made the cores learned to make the = cores=20 better, and to assemble them better, and at lower prices, it just opened = up=20 whole new opportunities for computers. In time we bought RCA's core = making=20 facility and the facility in Taiwan for stringing them. And when we were = going=20 at our peak, we were making 4 billion cores a month, and stringing them = in=20 Taiwan. At that time cores were cheap, memories were bigger than we ever = had=20 dreamed, and then we were able to make a large number of computers.=20

    DKA: Did the development of this machine, once you got the idea and = had the=20 core and decided to go ahead, did it go smoothly? Or did it have, was it = a rocky=20 road? How do you remember the development path?=20

    KO: The first one we built we bought standard cabinets from a = supplier. They=20 were rounded and kind of plain and ugly, with a separate console. It = became=20 obvious we had to make a better looking cabinet. One that was more = modular, we=20 could increase the size of and that would look unique and more colorful. = So, the=20 first one was really plain. But in time we came out with this design and = going=20 through that transition frustrated some people because it did take some = time.=20 But it probably took less than a year to build the machine. [SETTING IN = FRONT OF=20 THE PDP-8 COMPUTER]=20

    TO=20 CONTENTS

  • Designing the PDP-8

    KO: One time in the 60's, the Atomic Energy Research group in Chalk = River,=20 Canada, asked us to bid on a specially designed digital device to = control an=20 atomic pile. They gave us a whole list of equations they wanted built = into it.=20 We looked it over and decided we didn't have anybody with enough = patience to=20 design a special device to do those, solve those equations. We told them = that we=20 would sell them a very simple computer, at a price as if it were a = production=20 machine and that they could then program it to do these equations. They = liked=20 the idea because they said they didn't really know what the equations = were=20 anyway. We then, with that contract, designed our PDP-4 [PDP-5] which at = the=20 next go round was called the PDP-8. This machine was just 12 bits long, = the=20 shortest we could conceive of. It had the simplest order code and the = simplest=20 organization that we could think of. It was not designed to compute, it = was=20 designed to control something like... like a pile. Customers bought the = machine=20 and started doing programming with it. The great appeal it had was it = was=20 inexpensive. We did a number of bold things. First of all it was bold to = make=20 one that simple. The competition laughed at us. It didn't have the=20 characteristics of the computer. We also standardized a teletype printer = which=20 was not designed for continuous use. We very formally made the decision = that we=20 would gamble on that teletype and work hard to make it reliable enough = for=20 continuous use. That one gamble, and that one small success, probably = was a key=20 part in the introduction of minicomputers and personal computers. = [REFERRING TO=20 THE ASR-33 TELETYPEWRITER] Printers before that were very expensive. You = could=20 not have an inexpensive machine just because of the printer. That = machine was=20 quite an expensive, very cleverly designed but made for offices where it = was=20 used intermittently. Computer users are continuously at a very high = rate. And so=20 it was a very important development in the history of computers from our = point=20 of view. It was a bold thing. Because teletype [manufacturers] said, = don't do=20 it. It's not designed for this. In time they appreciated it and together = we made=20 it reliable enough for this use. People fell in love with that machine = because=20 of the cost. And because of the simplicity, it was easy to teach people. = Out of=20 that came very sophisticated software. It could safely be said that for = many=20 years that the PDP-1 was the one machine that introduced computing to = much of=20 the world who had any contact with computing. That's because it was = readily=20 available and inexpensive.=20

    DKA: Were you surprised at how successful?=20

    KO: Oh, yes. We were brave making a machine at all. We never were = brave=20 enough to think that it would become such a popular computer to do = computing.=20 The PDP-4 [PDP-5] was built like the old machines. When we were going to = redo it=20 and call it the PDP-8, we wanted to introduce new technology. Sylvania = had made=20 a new socket that IBM used and it took modules like this. [HOLDING FLIP = CHIP=20 MODULE] We went one better than IBM in that we used glass boards with = high=20 tolerance on the boards. We had 18 connections instead of 16. We had a = clever=20 looking handle that really became our trademark for many years. [HOLDING = A PDP-8=20 FLIP CHIP MODULE] We then had a set of technology that last[ed] us for = 15 years=20 or more. Maybe 20 years. It was something we just built on from there. = The=20 modules became quite long and quite high, in time. But for the PDP-8, = this was=20 the module. We also set about to use the technology that was developed = for home=20 appliances and automobiles. The technology we used before that was = basically=20 built upon military technology. So with this machine, we set about to = study all=20 the technology used for appliances. I went to several discount stores = for hours,=20 studying every single washing machine, dryer, stove, to see how they = built=20 things and what we could use. The [idea for the] switch handles came = from a=20 Maytag dryer. Using a glass panel without separate lights showing was = the=20 technology used in appliances. The silkscreen on the back looks quite=20 attractive. Inside we used the slip-on connectors that were commonly = used in=20 automobiles. We tried every way we could to take the technology that = made mass=20 production possible in appliances. The design of this has no screws = showing, it=20 is basically held together by these two lock switches. We always had a = lock on=20 our computers but only one. We obviously needed two in order to hold = this=20 together so we invented a use for the second switch. This one turned the = thing=20 off and on and this one disconnected the console so you couldn't mess up = things.=20 The fascinating thing is that after[wards] every computer had two = switches. Even=20 the Japanese computers had two switches. No one ever knew why there were = two=20 switches; it was to hold this panel together. But they just followed.=20

    DKA: When they looked at some of the home appliances, what was the = driving=20 reason?=20

    KO: The technology we used was expensive. There's very few inventions = that=20 come out of the blue completely. Most inventions come from adapting = ideas from=20 other people. This goes for ideas of organization, motivating people, = how you do=20 things mechanically, how you do things electrically, and just being = exposed to=20 all the things that people are learning in the appliance industry. = Opened up new=20 sets of technology to make things inexpensive. They mass produce things = at very=20 low price and we had to do the same. They also had more years of = industrial=20 design experience. They, every day, sold things on looks. And it was = more=20 important to us. So this was part of that. We had an industrial designer = who=20 said we should show off our new plastic handles by covering it with = smoked=20 glass, smoked plastic. I didn't like that idea but it turned out to be = great. It=20 helped sell. We had rosewood formica on the side which added a richness = to it.=20 This machine became the standard for industrial design for computers for = some=20 time. It opened up nicely on a hinge so you have access to the inside. = It gave=20 people the feeling it was put together with thought and easy to fix and = easy to=20 assemble. That machine then made a very important contribution to the = company=20 and to the industry. There's another story with respect to this that you = could=20 tell many times over. Our first PDP-4 [PDP-5] was laughed at by the = competition=20 because it was so simple. So naive. But in time, they learned to respect = it. We=20 knew they were coming out with an equivalent. We also knew that they = were making=20 the classic mistake that is made over and over again which is to look at = someone's old product and think you can do better than that. The thing = they've=20 forgot and they should have known better is that we also were working on = new=20 products. So one came out with their announcement, two weeks later the = other one=20 came out with theirs. And a week after that we announced this one at a = much=20 lower price and a much more elegant machine. That story happens so many = times in=20 business where people forget that their competition is not the old = machine but=20 the one the competition is working on. The computer is made up of = hundreds of=20 modules like this. [HOLDING A FLIP CHIP MODULE] They plug in from the = front side=20 into sockets which are assembled in the rack. And these sockets then are = automatically wired by a very large automatic machines. [REFERENCE TO=20 GARDNER-DENVER WIREWRAP MACHINE] So, all that wiring was put on = automatically=20 and it was one of the new technologies that were available at this time. = So this=20 panel we put into a machine and [could] take it out all wired. Put the=20 appropriate modules in the appropriate slots and assemble two of them = and we had=20 a computer. The one gap which maybe you can see is the memory itself and = the=20 circuits that drove it are also on these same modules. The modules = filled the=20 whole area except this one area which is core memory. [POINTING TO BOX = WITHIN]=20 It's got a fan on to keep it cool and is driven by circuits also on = these=20 modules and together it makes a whole computer. In the MIT tradition, we = had the=20 system divided into sections. With these switches we could separate one = section=20 and vary the voltage to see if any section was deteriorating and check = the=20 margins and replace it if it was in trouble. The power supplies were in = the=20 bottom. The whole unit was quite heavy. But as machines went, it was = very small=20 and very light. People then had opportunity to have a machine that was = all their=20 own and operate it in the same way we operate a personal computer today. =

    DKA: How is the business changing? Is the computer business a = different=20 business at this point?=20

    KO: In one sense the computer business stayed the same. It's just = that as=20 things got less expensive, more and more people were able to buy = equipment and=20 to use it. With a $110,000 machine you couldn't afford to give one to = everybody.=20 At $18,000 [THE PRICE OF THE PDP-8] you were getting close. As this = machine got=20 down to $12,000 and $8,000, it became possible to give it to every = technical=20 person to use. He'd could use it to do all sorts of operations.=20

    DKA: Who were your customers?=20

    KO: Schools were one important customer. Because with this you could = really=20 give students the opportunity to understand what a computer would really = do. It=20 was also very popular in controlling instruments and machine tools and = medical=20 devices. Anyplace where you had simple operations, this was just ideal.=20

    DKA: Are people starting to use your equipment by this stage for word = processing, or any type of business application?=20

    KO: We believe the first word processor was done on a PDP-1 at MIT by = the=20 students. They called it "Expensive Typewriter" because it would tie up = a whole=20 machine and an expensive electronic typewriter. It would do many of the=20 wonderful features we take for granted today. In time we offered word = processing=20 on every machine that people bought. The technical world had been using = word=20 processing long before the commercial market showed any interest in it. = The DEC=20 computers were the standard for word processing. We were slow in getting = into=20 the commercial area because the technical area kept us busy for a long = time.=20

    DKA: As your company was growing, Ken, in delivering more and more = products,=20 was the nature of the company changing?=20

    KO: The company changes consistently and regularly. The people we = hired=20 initially of course were not trained in computers. They came from all = kinds of=20 backgrounds. Musicologists was surprisingly popular. Then in time we = could hire=20 very well educated and trained people in computers. This made it change. = The=20 initial applications didn't need all the formality and discipline = because they=20 were small. The operator kept them in his head and he understood them = all. In=20 time, many people had to work on them. The applications had to last = longer than=20 individuals stayed on the job. This meant a lot of formality and = discipline in=20 the software and the hardware also. So the business changed over a = period, many=20 years from getting the fastest, most exciting thing out to supplying all = the=20 discipline necessary to make sure it worked forever. The question's = often asked=20 what part the military or the US government play[ed] in the development = of=20 computers in developing the advantage the US has to the computer market. = In=20 general, I would say they played little part. Obviously, Whirlwind was = financed=20 by the military. But during its development, it really didn't have an=20 application, so it wasn't driven by the military. The big drive came = because of=20 the high demand in the competitive situation in computers. The advantage = the=20 U.S. had was with a few exceptions. Not the fact that the government was = involved in the computers or financed things. But the fact that our = government=20 never could catch up and try to help us. Most other countries' = government=20 stepped in and tried to help the industry, and once the government did = that we=20 knew that that country was not in competition. Every time our government = tries=20 to step in, (and they're frustrated often because they can't run things = and they=20 can't tell us what to do), they mess up everything. Fortunately the = industry has=20 been running so fast they couldn't grab hold of things. When they tried = before=20 they did, they were left behind. I'm sure it's the goal of people to run = the=20 whole industry, and if it disappears and goes somewhere else in the = world,=20 they'd love to see that, too. But the biggest contribution our = government's made=20 is that they never got up with the computer industry in order to help. = No one=20 else except the Japanese have been successful. I think a big part of = that is=20 that their government's helped and they tried to be dependent on the = government=20 and that always sets them back.=20

    DKA: So that relates back to your philosophy of commercial = independence.=20

    KO: Yes. The military is always several years behind in computers, = and=20 getting farther and farther behind. They're not in a position to lead = the=20 computer industry just because of the way they're organized.=20

    DKA: Does that mean that you don't think your relationship with the = Japanese=20 government to that industry is a threat to our way?=20

    KO: The Japanese government appears to be the only government that = can help=20 industry. It might be because they don't consider industry an enemy like = western=20 governments do, particularly our government. But somehow, most of the = time they=20 do, they help. But I'm not sure that it's helped them to compete in = general.=20 Good commercial competition is what drives industry. We may be losing = that now,=20 unfortunately. We may lose it because the government wants to control = things. We=20 may lose it because the attitude today is stockholder's rights, they = call it,=20 where you take all the resources and give it to the stockholders = immediately and=20 invest nothing in the future and get rid of all the assets. [You] don't = invest=20 in research at all and that's the fad today. That's considered the = ethical,=20 American, moral way and if we continue that way we're sure going to lose = it to=20 anybody because there won't be anything left. But if we can avoid that = somehow,=20 we should be able to stay ahead.=20

    DKA: What's your view of the right recipe?=20

    KO: One has to believe in capitalism, not the rampant, stealing, = robber baron=20 type capitalism, but competition in order to do a good job and continue = to grow=20 and develop. When our government tries to resort to protectionism, or = thinks=20 they're going to control the world supply of memory chips, they just do = no end=20 of mischief. If they continue that way, you don't know they'll stop in = creating=20 mischief. If you believe in competition, you will buy from whatever part = of the=20 world does the best. And they'll buy from you where you do best. But if = you lose=20 one part to somewhere in the world because the cost of capital is less = there or=20 because they concentrated, because they worked harder, because they have = some=20 natural assets, if you try to change that by laws, you get up into an = impossible=20 situation and you'll destroy the advantages of competition. A politician = inherently doesn't believe in capitalism because it doesn't make good = politics.=20 Our biggest danger I think in technology is not the Japanese, it's our = own=20 politicians. Their claim to their constituency is they're going to = manipulate=20 everything. They just aren't able to do that in a wise way. All that = we've=20 gained we could lose very quickly by just following our trend of = protectionism=20 in the technical area, and also our trend in looking out for short term=20 interests.=20

    DKA: As an international company, does your worldwide reach protect = Digital=20 from that type of influence?=20

    KO: We do have the advantage of having access to every part of the = world. We=20 freely buy from any part of the world. We trust the Japanese. We have = operations=20 there. They are part of Digital. We'll trust any company or country that = makes=20 good products and is a reliable source, because we buy and sell = internationally.=20 This is what we say our country believes in, in theory, and it works = well and we=20 follow it.=20

    DKA: Going back to the PDP-8, was this one of the machines that = people then=20 began to network together? When did that notion get started?=20

    TO=20 CONTENTS

  • Networking PDP-8 Computers

    KO: One idea we took from MIT was the idea of networking. The = aerodefense=20 system was made up of 23 sites; each one having a very elaborate, very = high=20 performance local area network. There was a large number of display = terminals.=20 Then each of these 23 [sites] were networked to two arrays of radars = across=20 northern Canada. Airborne radars on the coast, human being spotters on = top of=20 the tall buildings. It was one very large network. So networking became = a theme=20 in everything we did. We always were able to hook our computers together = and for=20 almost all our history we were able to hook to IBM. Networking was = always part=20 of it. The big change in networking came 15 years ago or so when we = decided that=20 we'd have to network in a very standardized way. Everything we did we = used the=20 same networking protocols, and the same networking technologies. That = really=20 made networking a major part of our organization. It's these = standardized ways=20 that we've been encouraging the world to accept so that we and they can = all work=20 together on the same network. We're well under way in that and the new = standards=20 will make that possible.=20

    DKA: Ken, were you unique in the industry pushing for that type of=20 discipline?=20

    KO: I think we were quite unique. There's a normal tendency to keep=20 everything you do secret so that others can't come into your area, just = like the=20 railroads in Europe used to have different gauge railroad tracks. The = reason=20 they did that, was that if someone was invading them, they couldn't = drive their=20 trains into your country without stopping to reload. This created an = enormous=20 bottleneck for any invader. People tried to maintain their own standards = so that=20 anybody coming into their customers would have a terrible time. We had a = different idea. It was clear to us that we never could own every = computer a=20 company had. It was clear that ideally everybody's computer would be on = the same=20 network. So we've been pressing for standards to make this possible for = along=20 time. The standards that we've had with great discipline within our own = company=20 means that we can tie anything we make together, anywhere in the world = quite=20 easy and simply and elegantly. As other people accept the standards, we = can=20 accept them in the network. Eventually if all the plans now follow = through, the=20 major companies will all accept the same standards.=20

    DKA: How are all the companies reacting to this?=20

    KO: They all have great plans. Some of the standards were basically = ours,=20 some of them were IBM's, and some of them were generated independently. = There=20 seems to be common agreement that once they're accepted, everybody will = follow=20 them, independent of where they came from.=20

    DKA: Thinking about networking on one hand and the type of hardware=20 development that goes along with that, it seems that you followed a = pattern in=20 Digital starting with the modules, [and] state of the art memory systems = that=20 continued through the PDP-8. How did that hardware pattern continue into = your=20 later machines? Or did it?=20

    KO: The hardware continued to evolve as technology evolved. If you = look at=20 what we do today, it bears no relationship to the early modules. But = step at a=20 time you can see the evolution was quite clear. Today we tend to build = things on=20 large boards with large integrated circuits, and the amount of handwork=20 decreases every year. In the last six years I think we doubled our size = twice,=20 approximately, and we increased a number of people in production by = almost zero.=20 This comes about because our way of building things gets to be more and = more=20 efficient, and more and more gets built into the integrated circuits. If = you=20 look at it that way, things have changed dramatically. But if you look = at each=20 step, each step was a small one.=20

    DKA: What resemblance do integrated circuits bear in terms of logic = to the=20 earlier modules?=20

    KO: At first, integrated circuits would have maybe the equivalent of = this.=20 Now the integrated circuit has the equivalent of ten of these. I don't = know,=20 maybe it's a hundred of these. Because they're so cheap they're used=20 differently. We had to be very economical on how we used these because = they're=20 expensive and took a lot of room. The logic is so cheap with integrated = circuits=20 that you use it freely and do things that are inexpensive because you = use so=20 much of it. So it's a little hard to weigh the differences. But it's = created a=20 revolution, a slow revolution because they started off small and now = they=20 include so much. But as they get bigger we want to do more. As they get = bigger=20 we have the problem of cooling them, connecting them together and the = technology=20 puts new demands on how we do things. We even cool some of them with = water=20 because they get so much concentrated on one area that takes different = ways of=20 cooling them. So things are, over a period of years, quite different. = Even=20 though any one year they don't change very fast.=20

    DKA: And as you move from the PDP-8 to the PDP-11 family, was that a = major=20 shift in your computer company?=20

    TO=20 CONTENTS

  • Developing the PDP-11

    KO: The PDP-11 was a machine we took a long time to design, and = worked hard=20 to make it one that would last a long time and have innovation in it = that would=20 make it unique for a long period of time. It's still a major product for = us, and=20 it must be 15 or more years old. We set about to continuously improve it = but=20 still keep it an 11. We had many software systems, each one did unique = things.=20 One of them was the predecessor for the common PC software today. Others = did=20 other things especially well. The difference in the 11 is that it became = the=20 center of the corporate strategy and the resources were all put on it = and yet it=20 was maintained with discipline so that the same software played on each = machine.=20

    DKA: When you say it became the center for corporate strategy, what = does that=20 mean?=20

    KO: Before that, we had several computers going at one time. Almost = all our=20 work was concentrated on that one. We also had a very large machine, = called the=20 PDP-10. But the small minicomputer area was concentrated on that 11. The = next go=20 around with computers, we carry this to a farther degree. The VAX = computer was=20 done with a lot of planning before it was started. It was planned so = that=20 eventually it would span a range of sizes from very tiny. It was = designed so=20 that the same software written ten years ago would play today and ten = years from=20 now. It was designed so that any part of it could be taken out, = improved, redone=20 and a new one put in and it wouldn't upset the whole system. It was = designed=20 with just one software system. It also did UNIX, ten percent of our = machines=20 were sold to do UNIX, but the basic software system was only one called = VMS.=20 This is obviously very productive, everybody works on it. It gets better = and=20 better every year. If the whole company works on one [operating system], = it gets=20 to be quite good after a period of ten, fourteen years. But it is = contrary to=20 the normal inclination of engineers. With our PDP-11, and like most = other=20 companies, the tendency is to say, with a new software system I can make = something special. It does things so much better. Having discipline to = avoid=20 that made everything so much better just because we could concentrate on = the one=20 thing we were doing. That's what made the VAX very popular.=20

    DKA: And that was a discipline that was hard to maintain in your = company?=20

    KO: It was easy for me because someone else did it. Our strategy said = we=20 would have one protocol for networking which the idea is somewhat unique = to be,=20 have only one. One computer architecture which was the VAX, one software = system=20 which is the VMS, and one way of doing local area networking which is = Ethernet.=20 The discipline came automatically. Parts of it I really had to take part = in=20 because without corporate discipline, people would use different kinds = of=20 networking, and you would lose the whole strategy. So it was not my job = to=20 develop the strategy. People did that very well without me. Even forcing = it=20 through and getting corporate approval wasn't my job. But in time it was = clearly=20 my job to make sure we followed it. In general people liked to work with = a clear=20 strategy. They like to know where the goals are. The basic decision on = projects=20 are already made. That gives them great freedom to be creative in the = areas=20 where we need creativity. Not everybody likes it. When we finally got = around to=20 enforcing the strategy we had all along, a large number of the vice = presidents=20 quit because they wanted freedom to do what they wanted to do. But the = result of=20 the whole company working in one direction was obviously very good.=20

    TO=20 CONTENTS

  • The Role of the Personal Computer

    DKA: Most people think about computers as having fundamentally = changed. When=20 the personal computer came, truly to be on their desks at home, at work, = how did=20 the advent of the microcomputer effect your business? And how did you = see it and=20 how did you respond to it?=20

    KO: The goal we set about when we started the company was to = introduce=20 interactive computing. We did that first with the PDP-1. Then with = timesharing,=20 where one computer has many terminals and each person thinks he has a = computer=20 to himself. With the timesharing end of the terminals, people then were = able to=20 do those things which we see with personal computers today. As they did = word=20 processing, they did computation. They eventually did spreadsheets and = many, and=20 did games and other things. At one time at my home we used to play = Scrabble on a=20 terminal plugged into the telephone with a VAX at Digital. When you had = a=20 terminal you had everything that we see now on a personal computer. So = that's=20 been around for many years. We saw in the early 70's that it was going = to be=20 easy for people to make computers. The type of computers, we had made = more=20 powerful than this one, were going to be able to be made by anybody very = simply=20 and very cheaply. At that time we set about to do something more = difficult which=20 was to integrate or network a whole organization around the world = together.=20 Within one room, within the building, within a campus, a city or the = world. And=20 that was our thrust. We concentrated on that because the PC, as it was = being=20 developed, was so easy there'd be many people making it. So, we in = general=20 avoided it and [instead concentrated] on the problem of networking them = but not=20 planning to be the large producer of PC's. We concentrated on the making = of=20 networks of small computers and large computers which is a much more = challenging=20 job and devoured all the resources we had. The PC itself was a component = to the=20 network. We made some PC's designed to be part of the networking but the = general=20 PC market was not for us. There were too many people in it and it turned = out to=20 be true. At one time I think there was 500 or 700 people making PC's. = Anybody=20 could build them. You could build them in your basement. That was not = for us.=20

    DKA: Was there doubt about that decision? Or debate about that = decision?=20

    KO: No, because you see our goals were clear and when anybody can do = it and=20 there's nothing particularly unique that we can contribute, it's clear = it's not=20 for us. Now we had PC's demonstrated here long, probably long before = anybody=20 else did. Individually people would make them. But we very formally = decided that=20 was not what we were going to do. It would basically be a very good = decision.=20 The IBM success in that business was, for a number of reasons, partly=20 happenstance, partly luck, but to a large degree because they had the = size, the=20 resources and the experience to set up the infrastructure to deliver = millions of=20 these. It was not a matter of invention, it was a matter of management = and=20 resources. They were the ones who could do it. After they had done it, = it became=20 easier for others to enter the market. But their contribution was good,=20 competent management. And we were off doing other things. Now the = argument we=20 have today is an interesting one. We believe in PC's. We encourage them. = We=20 network them. We use them in large numbers. But we still believe that = most=20 people in an organization want terminals. Terminals you don't have to = worry=20 about data management, you don't have to worry about floppy disks. You = just sit=20 down and it does the work for you automatically. So our most = experienced,=20 educated computer scientists and my secretary who has access and = experience with=20 everything, always want a terminal. It's just so simple to use. There's = nothing=20 there. And the secretary doesn't want to take her hand off the keyboard = and run=20 a mouse, so the terminals we feel will always be important. That's a = major part=20 of our business. We do supply PC's, we will supply more PC's, we = integrate and=20 work with other suppliers, all the suppliers of PC's, to network them. = They have=20 a very key part in this system. Now there's the large expensive, very = competent=20 PC called a workstation, offering a whole new realm of things. That's a = very=20 important market. These cost twice as much as PC's. They have a very = beautiful=20 display, sometimes with beautiful colors, always with fine detail, and = are used=20 for designing automobiles and airplanes, and for many things for which = you want=20 very precise pictures or a lot of material on this display. They get to = be=20 powerful when they're hooked up to networks. Now there's an interesting = mistake=20 that people make, the press makes out of innocence. Six years ago they = announced=20 that the PC's in a operation have enough computer power to replace the = big=20 machine. That turned out to be foolishness. Now they're announcing that = PC's,=20 that workstations networked together, have more power than a big = computer, and=20 they'll replace the big computers. That's nonsense too. They replaced = some of=20 the things big computers were doing because they can do the jobs, but = they don't=20 replace the big ones because in any organization you have data that you = cannot=20 afford to lose under any circumstances. The last thing you're going to = do is=20 have your key data in somebody's workstation where somebody can mess it = up. All=20 the protection for precise data has to be separated in a place that can = never be=20 lost, never be damaged. And no way are you going to leave it out in the = open on=20 a small machine. So there's a place for everything. The PC's will play = part of=20 it, terminals another part of it, workstations another part of it, = medium=20 computers and large ones other parts of it. There's a place for all of = it.=20

    DKA: You've been in a relatively unique position for the past three = decades,=20 four decades maybe, to watch this industry, and to think about where = it's going.=20 What really stands out as the key developments in a large social sense = in the=20 elevation of the computer and the way it's effecting our society. What = do you=20 see as some of it?=20

    TO=20 CONTENTS

  • Observations About the Computer Industry

    KO: I think the interesting thing to observe about computers and = computer=20 technology is that the most significant changes people don't notice. = Things they=20 worry about never become a problem. For example, the hand calculator = really was=20 a revolution. No one predicted it, no one worried about it. It sneaked = up on us=20 and suddenly we all have them, we all use them, and we never thought of = it as=20 revolution. It just sneaked up on us. All through our life there are = computers.=20 Our cars, our homes have computers, and so many things, and we don't = notice it.=20 We worried about privacy, worried about computers running our lives, and = those=20 didn't happen. The issue of privacy is that with computers you can = specify,=20 society can decide exactly what level of privacy we want. You can have = anything=20 you want, which you didn't have before. You're going to run your private = finances with [the personal computer]. They're going to run your menu. = They're=20 going to run your social affairs...didn't happen. You don't want it and = it=20 didn't work. The things we feared usually don't happen. The things of=20 significance sneak up on us and we take them for granted like they = always were=20 there. When I was young in the Boston area, you had one charge card in = the big=20 department store in town. Always in trouble. They never got anything = right. And=20 always patching up something. Now you couldn't carry all your credit = cards, and=20 very rarely do you have trouble with them. That's a revolution. Now we = expect=20 not to have trouble with our credit cards. We don't expect them to make=20 mistakes. That's the computer revolution that we just take for granted. = And it=20 goes on and on and on. We have a long way to go. When we ever straighten = out=20 medical billing, it will be computers that do it. And after it's = straightened=20 out, we'll never remember that it wasn't straightened out. There's two = things I=20 would say to people about computers. One is, don't fear them. Most of = the time=20 they're doing good for you. The other thing is, don't ever become lazy. = Remember=20 that you only have fun in life and you only can stay ahead if you keep = learning.=20 Calculators are not an excuse for learning how to do arithmetic. As for = things=20 that computers do, you've always got to be sure that you can do a = certain amount=20 of things by hand and know something and learn something. Always learn, = and=20 don't ever let computers fool you into thinking you don't have to. I = tell our=20 people when I'm asked to lecture, look at the old people you want to be = like. I=20 can tell you ahead of time that if people continue to learn, are excited = and=20 know about things, and the ones who are boring are those who stop = learning and=20 don't think about things. Just don't let computers cause you to get in = the trap=20 of not losing things.=20

    DKA: What's coming?=20

    TO=20 CONTENTS

  • Future Computer Inventions

    KO: From a technological point of view, I think we can be confident = that=20 computers are getting more powerful, and less expensive. This means that = we can=20 do things we never thought of doing before. It means that people can use = techniques for doing computation which are wasteful, devouring memory = and=20 devouring computation and doing things in a way which is so much easier = because=20 the computer does all the work. But what it means to the private = citizen? It=20 means that your automobile will run better, your house should run = better, and it=20 means that we can, little by little, get better service from all the = things we=20 struggle with today. And for the people who really have a use for the = computers,=20 they'll do things a much more exciting way. It also means that every = student=20 should learn to type. Anyone who does any writing at all should have a = word=20 processor. It means that people shouldn't be afraid of computers even if = they=20 don't have any need to learn. But anyone who has any reason to learn, in = all=20 young students, really are in that category. They should have a feeling = for=20 computers. It will just continue to be a contribution to society.=20

    DKA: Does anything scare you about what's coming ahead?=20

    KO: I used to think that computers could do no harm. But there are = some=20 things which do worry me. Some people study computers and don't learn = anything=20 else. Computers are just tools to do something; you better be expert in=20 something and consider the computers the tool. The computers are fun and = exciting but they're just tools, and we better make sure that we know = something=20 about what the computers are supposed to solve not just the computers. = Computers=20 also produce an enormous amount of data and people get confused with = that. Data=20 is not information. That's been pointed out. You put the data in a form = which is=20 useful and you have some information. But a large amount of data isn't=20 information. I think in business making graphs is a menace. Very few = people know=20 how to make graphs. They don't know why they make graphs. They make = graphs=20 because they think you're supposed to make graphs, but they don't know = what=20 they're trying to get across. They don't know the reason for it, and if = they had=20 the reason, most of the time they wouldn't do it. In that way, computers = are a=20 menace to business. We have more graphs that mislead or cause confusion = than we=20 ever did before. But, in general, I think computers do a lot of good. We = have=20 too much information, but it's so much better to have free flow of = information=20 even though it turns out to be a little too much than limiting = information.=20 Computers are going to revolutionize business and society, even maybe in = particularly the closed societies because with a free flow of = information, it=20 just changes the way we do things. Looking at the big old machines, = there's=20 almost no relationship to, or no indication of what a computer will do.=20 Computers will do almost anything you want. If we can't afford to do it = now, it=20 can [be done] in the future. The one thing to learn about computers is = that they=20 do give us the opportunity to accomplish things we've never been able to = do=20 before. This means that our big problem is we've got to decide what we = want to=20 do. Most people don't know what they want. Often in society we don't = know what=20 we want. But if we decide what we want, most of the time the computer = will play=20 a key part in giving it to us.=20

    TO=20 CONTENTS