An Interview with LaRoy Tymes

LaRoy Tymes


LT = LaRoy Tymes
GAM = George Michael




GAM: Today is July 1, 2006 and we are going to re-interview LaRoy Tymes; the Laboratory's Technical Information Division having lost the first tape. So LaRoy, please begin by telling us how you got to the Lab and we can go from there.

LT: OK, before we even do that, though, I've got to tell you it's a strange sensation coming in here seeing you, I feel like I'm a kid again viewing the old master. It's a strong feeling, and it's really odd because, at sixty three years old, I'm not a kid anymore, but that's the way I feel.

GAM: I'm eighty. So you're right; I am old, but you're the real master.

LT: So anyway, I was in Berkeley and I was all set to become an apprentice electrician. I heard they needed computer operators up on the hill, so I went up the hill, had an interview and they told me about Livermore. I'd only been in California very shortly. I'd never heard of Livermore before. They said I'd probably be more suited at Livermore than at Berkeley, so they told me how to get to the Lab. The next day, I drove out to the Lab. Highway 50 was just a two-lane country road, and it was over a hundred degrees. Here I was in the middle of nowhere. I met some people at the Lab, and they said "don't call us, we'll call you". I went back again, still determined to become an electrician and I got this letter in the mail saying "come on out, we'll hire you." I was too na�ve to realize that they would start paying me to do a "do nothing" job even though I didn't have my clearance yet. So, I got a job making pizzas and did a number of other odd ball things. Finally, I came in there and sat a few weeks in the cooler and taught myself how to do FORTRAN. Then I went to a water-cooled reactor, but still didn't have my clearance, and I got to do a number of neat things, not operating the reactor itself, they had licensed operators to do that, but I got pretty good at operating a number of the peripherals of the reactor and a lot of its equipment. I got to go into the reactor room and do a number of experiments.

GAM: Just for the record, what year was this when you started?

LT: I started in 1963. I believe it was November or December, actually. I got to Livermore the summer of 1963, in August.

GAM: OK.

LT: There was a guy working on his PhD. Thompson, I think his name was, had me counting indium foils. He would expose very thin foils made of indium to plutonium at different distances in the water tank for certain amounts of time. It had something to do with computing a Fermi number. I never really understood the physics or mathematics behind it, but I did that for awhile until my clearance came through.

I finally got to see the computer center, and I was astonished at the rows of consoles and computer tapes and all the other equipment. It was way beyond anything I'd ever seen in my life or even heard of. I had no idea that such a place even existed. I was also very impressed by the fact that they would entrust these multimillion dollar machines to somebody like me. Nobody had ever given me that level of responsibility before. These are machines where you had to account for every minute of operation, because every minute was precious. You couldn't let it sit idle, you couldn't let one job run and interfere with the scheduling of another job. The schedule was all important. You got to be very good at changing tapes very fast for instance, because if the machine was waiting for a tape, this clock was ticking at several dollars a minute, which in those days was a lot of money. So I got to learn to change tapes, these ten-inch reels on the IBM and Control Data tape units.

There was one guy on the shift, Gary Tollefson, who I was always competing with, but he had more years of experience than I had and he was always just a little bit faster than I was.

We had calluses on our hands because one of the jobs the swing shift operators had to do at the beginning of every shift is called "tape scripping." The first few feet of the tape were worn more than the rest of the tape, so they would have bare spots on them so you had to cut those off and then put a low point reflective, little foil load point a few feet up from the leader of the tape and then wind that back up again. To do this operation required a lot of twirling of the tape on your hand, on your left hand, so your left hand always had a characteristic callus. If you were a Sherlock Holmes you would have no trouble spotting the operators on the swing shift because they all had that callus on their left hands.

Being a swing shift operator, I learned fairly quickly that if I was doing chores like getting tapes or scripping tapes or something that took me away from the machine I was responsible for, I would have a cheap transistor radio sitting next to the intercom tuned between stations. It would pick up the radiation from that machine. 7094's radiated especially strongly. After a while, I learned to tell by the noise that was picked up by the radio which program was running; if I wanted to check on your machine without being in the room, I just checked the intercom, listened to my particular radio, which was near another intercom and if it was playing the wrong music or it wasn't playing anything at all, I knew I'd need to run back to my machine because, something had happened.

You were asking about the tape program, the Stars and Stripes Forever, on the Control Data 3600. I had made it a goal for myself to learn how to write machine code for every machine in the Lab. I never succeeded with the LARC, because that machine went away before I ever learned to code it. But, for all the other machines, including the 650, which was just about to be shipped off because it was obsolete at that time. I wrote some hand code program or another to learn the architecture of that machine. When I did the 3600, I decided to write music for it. I'd already written music for the PDP-8 that you could play back by sitting an AM radio next to it and just listening to the radiation from it. The 3600 had very high performance control data tape units that could read and write 200 inches a second. And they had spinning capstans that spun in opposite directions and had inner buffers so that if the tape wasn't moving the capstans were separated from the tape by a very thin layer of air. And then if a capstan wanted to grab the tape the air pressure would change to a vacuum allowing that capstan to grab the tape and accelerate it. The capstan was already up to full speed and as a result of that it would start and stop the tape very fast. The inter-record gap between the records was 6/10 of an inch. So you had 3/10 of an inch to accelerate to 200 inches a second, read or write the record and then 3/10 of an inch to bring the tape to a stop again once you got past the record. I found that if I started to do a tape write, and then aborted the write before the write began, I could start and stop the tape up to 300 times a second. 300 times a second was in the audio range and the tape loops went into their vacuum columns so the columns had a loop of tape in the bottom that was jumping up and down at that frequency, it made a good sound box. Especially if you left the front door of the tape unit open.

Then the next trick was to dial several tape units to the same logical number. So that they would start and stop synchronously, and doing all those tricks I could get a lot of noise out of those tape units. And I could control the frequency more precisely. Middle C is 256 cycles a second so I could play higher than middle C if I wanted to. So at the beginning of Stars and Stripes Forever when those tape units were going "Baum, Bop, Uppabaum", you could feel it in your chest. Even the building would shake, and for people who have never seen a computer behave that way, it was very impressive.

GAM: It was an impressive thing, very impressive.

LT: It caught almost everyone by surprise. You could tell them, "Oh, I got a music program for the 3600" and they would say "Oh, yeah, that's cute", but they wouldn't take it seriously and then they walked into the room and then suddenly it started to perform and everybody was astonished.

GAM:

Right. Well, there was some agitation in Sid's office to somehow punish you, but family day came along soon enough and your music program was the star attraction on family day. Everybody was just blown away by the thing. And it was decided that since it was such an impressive thing the guy who wrote it had to be impressive and we're going to keep him. We had to have him. So you were a hero, see?

LT: From villain to hero in one step.

GAM: Well, I think, to me, it's the most impressive demonstration of real time programming I've ever seen. There was a guy by the name of Peter Sampson, who for his thesis at MIT developed a music notation for the PDP-1 and one could write the music in voices. It would compile up to eleven of them and would play any three. But truthfully, your program was a much better player. So Bach's trio sonatas were Peter's choice and I kick myself, I had all those tapes and I was forced to throw them away as part of an office relaxation. I couldn't keep everything and I'm sorry about that. I'm even sorrier that I didn't keep a copy of your "Stars and Strtipes" program either.

LT: Well, none of us had very much foresight, myself in particular.

GAM: Yes, I understand.

LT: I never thought to record it or capture it in any way. I wish very much I could find a working Control Data 3600 today. That program became a standard part of the Control Data maintenance tape. So it went around the world, every field engineer had a copy.

GAM: Wish I'd known this just a couple of weeks ago. Bob Price, the last president of CDC, was out at the museum giving a talk. He's a friend of mine and I could have found out from him where to go. Are there any 3600's around? I'm not sure what we have at the museum.

LT: Where is the museum?

GAM: In Mountain View. It's the old graphics building, on Shoreline Boulevard, right next to the freeway. And it's got the world's largest collection of hardware from this era. You should make it a point of visiting sometime. You'll find that makes you feel young too. And they're doing a nice thing you know, they've got the hardware collected and it's going to set up in working displays wherever possible. Also there's a great big effort in software there too. It might, indeed, be that you could get a 3600 running there. They have a PDP-1 running there and also an IBM-1620 and a 1401. So a 3600 can't be that difficult.

LT: A 3400 would do too because I ported it to the 3400.

GAM: Yeah. Well, I'll take as a personal effort to go there and see what I can find out.

LT: Another comment on that villain/hero dichotomy, I ported the program to a 3400 and instantly burned out the tape controller. And I thought I was in deep doo-doo because that cost a lot more money than I had and I could not possibly afford pay for that kind of damage. And the field engineers instead of being mad at me were pleased to find that I'd found a flaw. It was a design error and they wanted the design error found and fixed. So it was, it was changed so that it could play my program.

GAM: That's good.

LT: Later on I tried porting it to a 3800, but I never got it work quite right on a 3800 because that had a cache, the cache messed up my instruction timing and I never quite understood how to get around the cache. I know a lot more about that sort of thing now than I did then because I was still learning about that stuff, so maybe now I could do it, but at the time I couldn't.

The Control Data 6600 was another failure because I couldn't control the tape units directly, it didn't have a speaker and it didn't radiate at all. I sometimes wonder if not radiating was actually a requirement, because you could tell from a considerable distance from a 7094 what physics program was running on it. Which made me wonder if people driving by on the road outside the secure area could listen in on it? I don't know if anybody ever did or not, but from the 6600 you couldn't tell even if you were right next to the machine.

GAM: Yeah. Well, but that wouldn't have anything to do with word lengths.

LT: No, it had to do with un-terminated cables running around; on the 6600 the high performance stuff was inside a cabinet.

GAM: Well, there's some critical distances there that they had to make sure were honored. And even more so in the 7600.

LT: Yeah, both of those machines, they had a number of twisted pair connections where the length of the connection which determined the propagation delay to get from here to there had to be tuned. Had to be the correct length. Not just long enough, that wasn't good enough; sometimes you added extra length to add a little extra delay to resolve the race conditions. So the architecture of the 6600 had a central processing unit and ten peripheral processing units. At least to the programmer it looked like there were ten peripheral processing units. I suspect in hardware it was really one box that was pipelined.

GAM: Right, that was called a barrel I think.

LT: Yeah. And why ten instead of a power of two, I don't know, but Seymour Cray probably liked ten better. It also seemed, in hindsight it seemed odd to me that the word length was sixty bits instead of sixty-four.

GAM: Yeah, I don't understand that either, but, I did hear an explanation though I don't remember it exactly. Seymour said it had to do with the availability of the memory units. So it was probably 48 bits with some extra bits on it for ECC and stuff like that, but I don't remember.

LT: Another factor is the 160 was one of Seymour's first machines. It was a twelve bit machine.

LT: Twelve bit machine. And sixty is a multiple of twelve, I don't know if that is relevant or not.

GAM: It probably is.

LT: The peripheral processing units were twelve bits.

GAM: You know, after Seymour graduated from college one of the first things that he did was to work with ERA. The Navy was funding them to build those code analysis machines and a lot of that was decimal. And that may also have had something to do with 60 bits, but I don't remember. But, go on with �

LT: OK, the character length was six bits, so there were ten characters per word. And one of the main programs I was responsible for was APT, (Automatically Programmed Tools.) That was originally developed on the 709, ported to the 7090 and the 7094 and I originally ran it exclusively on the 7094. Control Data ported that first to the 3600 and then to the 6600. The 3600 was forty-eight bits so there are eight characters per word and when they moved it to the 6600 at ten characters per word they didn't make a few changes that they should have to accommodate the difference in the word lengths, which added some conversion bugs. There was also a bad case where they typed in the value for Pi and reversed the last two digits when they extended the word length to the 6600. There were those and quite a number of other bugs that I reported to Control Data and they never fixed them, because to fix them, they needed clearance from Minneapolis. The group doing it was in Palo Alto and

GAM: It was Sunnyvale, really, right?

LT: Sunnyvale? No, it I think it was Palo Alto, this particular group. Dorothy Hornig, I think was the leader of that group. But, I was absolutely astonished that I'd gone through the work of finding and fixing these bugs and giving them my code for free, why didn't they just plug it in. And I was not familiar with their political constraints.

One of the more astonishing things was that the 6600 was slower by a factor of about a hundred, than the 7094, at running APT. And the reason was the 7094 was extremely dependent on its tape units. It had lots of tape units and used them all very heavily. The designers of APT for the 7094 worried about things like if they want to get at a particular record on a tape unit is it faster to backspace a certain number of times or is it faster to rewind the tape and re-forward. And they had all of that highly optimized to make the tape units work as efficiently as possible. When Control Data ported it to the 6600, they ported all that stuff over on simulated tapes, where it made absolutely no sense whatsoever. So if you did a backspace to a simulated tape, the tape was simulated on these big Bryant disc drives. So the big Bryant disc drives, to simulate a backspace, would go back to the beginning of the simulated tape and read-forward one record at a time until they were one record prior to where they had started. So if you wanted to do a hundred backspaces it might take many thousands of head motions to accomplish that task. So I spent about three or four weeks taking that part of the APT system and re-writing for disc. And by treating it as a disc and taking advantage of all the memory that the 6600 had I knew very often that if I wanted a particular record I could either find it already in memory or if it wasn't in memory I had a pretty good idea of where it was on the disc and I could go right to it. And that simple change made the Control Data 6600 run APT faster by a factor of a thousand. It made a huge difference, but Control Data never implemented it because that would have made it non-compatible with original 7094 version.

GAM: See the raw speed comparisons were a factor of six between the 7094 and 6600. Something like a factor of six. But getting a thousand is very impressive.

LT: Well, to get huge gains in optimizing a program the big trip is to start something that is truly awful to begin with. And that APT system was really, really, awful. That wasn't the only thing wrong with it; there were quite a number of things wrong with it.

GAM: OK.

LT: Another thing probably worth mentioning here is when I got my Batchelor's degree, I think in 1966, and then I became promoted to Programmer. And as a programmer, a guy named Terry Allison had been responsible for supporting the machine shops and the drafting department and anybody else who was interested in making mechanical drawings or getting things from some idea in a Physicist's head into some physical object. And Bud Wershing was the manager of that department, he didn't have anybody to replace him, so I came along and I was more than willing to do the job because that got me into a lot of different areas that I found fascinating.

I got cleared to go into practically every area in the Lab. There was no part of the Lab that I didn't have some curiosity about and I was especially fascinated with the machine tools. They had some pretty exotic equipment there; in fact, they had a motto, "why use lead when gold will do." And that attitude pervaded everything. So, because of that, I was able to do some pretty far out stuff. One of the problems that Allison faced and he didn't quite know how to handle was that a lot of the programs he was running were running on old machines that were obsolete like the 650 that was leaving the Lab, in fact, it had already left. And he wasn't facile enough to reprogram all of that stuff. I had to reprogram all of it in FORTRAN, I chose FORTRAN II for the 7094, but now I had the problem of instead of duplicating the old programs exactly, I found a lot of these programs did more or less the same thing. And I thought it would be a lot easier if I had one set of programs that did one thing and one way. And that would have the side benefit of getting all these different groups of people who were involved in a particular process to have all the same input and output and speak the same language. I thought it was a good idea but I ran into very strong resistance in some areas because that meant changing what people were used to. One particular individual, I thought, was going to have a heart attack; he was screaming at me and said he was going to do all sorts of terrible things to me. But I knew he was in a different Division of the Lab than I was, he could not do anything to me directly. He'd have to complain to his boss who'd complain to his boss and so on until eventually it would get back down to the other side in Computations Division, to Bud Wirshing, who didn't care. So I knew there was nothing he could do about it, he was going to take whatever I did and that was that. So indirectly, I think, probably, my biggest contribution to the Lab, or to the Atomic Energy Commission, was I standardized a lot of things that should have been standardized long before I arrived on the scene. It wasn't just the Lab itself that was involved, there were other parts of the AEC, in particular, Oak Ridge, because they would often receive the output of what I had done.

GAM: Interestingly, today, Oak Ridge is sort of like Livermore used to be. Their getting new machines and a lot of "gung-ho" stuff going on there, and the Lab is getting, I don't know, "long in the tooth" as we say. It doesn't seem that the Computation Department is just not as exciting as it used to be. I may be wrong.

LT: Well, when I was there the machine shops that I was supporting were the very best in the entire world. They were always doing things that nobody had ever done before, working with materials that nobody had ever fabricated before.

GAM: Well, that was true for lots of groups. They were doing things that people said they couldn't do and things like that. That was a great pleasure, but now all that's been disassembled.

LT: I found out that uranium was a very nasty thing for machine shops, because first of all, it burned rather easily like magnesium, and once it started to burn all you could do was throw it in the sand pit and cover it up, because if you threw it in the water it would burn even faster. Another problem with it, was that after you made a casting of uranium and then started to machine it, it had internal stresses that had to be relieved, so you had to leave some excess material on your part and put it in a hydrogen furnace and anneal it for several hours, put it back on the machine that you started with and try to index back to the original position as well as you could to make the final machining cut. And it was a very tedious and difficult material to work with.

Plutonium was another big problem because they didn't care that it was radioactive, what they cared about was that breathing Plutonium dust was very toxic. So that ought to be done in a hooded environment and you had to wear funny suits and stuff like that, because you didn't want any Plutonium dust to get on you, or especially in your lungs.

The only material that gave them no problems and that they didn't mind was the plastic explosives. The plastic explosives machined very nicely and were easy to handle.

GAM: I didn't know they machined plastic explosives.

LT: They machined it to extremely fine tolerances. In fact, some parts had a number of exacting requirements, where they had to be machined to very, very fine tolerances. But, that's one of the reasons I'm not too worried about terrorists getting materials and making certain things because even today, they are very difficult to build.

GAM: Well, I don't think we want to go further down that path.

LT: OK.

GAM: We'll stay on an unclassified level.

LT: Alright. So, anyway, I consolidated the entire group of programs so that they would start with some physicist having an idea of something he wanted to do and it would trickle down a chain of people finally to a draftsman who would have to draft it up and a machinist would have to machine it. It would go to the metrology department where it would have to be inspected, finally assembled and my software was common to all of these groups, which made it much more convenient.

GAM: They could talk to each other.

LT: So that they could all talk to each other, use the same numbers that meant the same thing.

Along that line, this is going back into the APT world. I wrote some of the first software to allow machine tools like milling machines to follow arbitrary curves. These splines that later became important in the aircraft industry because machining propellers and other kinds of air foils don't follow any mathematical formula, they're sort of a free form, or that's best represented by a normalized cubits blind curve. And nobody had the machine tools or the software to program a machine tool to machine such a curve, until I developed that at Livermore.

GAM: This stuff is still in use, isn't it?

LT: Well, it's commonplace, now everybody uses it. At that time a computer controlled machine tool was science fiction. Most people couldn't imagine such a thing. Now days, every machine shop has computer controlled machine tools. It's the only way to go.

GAM: I think I'm going to change the tape. OK, we're back on the air.

LT: OK it's going.

GAM: Yeah, well, we can sort of back count for a little bit. You left the Lab in when?

LT: I left the Lab in 1967, to finish off my masters degree at Cal State Hayward. I Went to Time Share in November of 1968, no, I'm sorry, in February of 1968. Norm Hardy and Ann Hardy had both left Livermore at that time and they were both working at Time Share and they invited me to come over. Norm still had a security clearance at the Lab and both of us went back to the Lab from time to time, but I didn't do any serious work at the Lab after that.

GAM: Ok, but you did do some fairly seminal work with Time Net.

LT: Yeah, Time Net from ah, I started on that at February of 1968 and then November of 1971 I brought up the network in all its glory and it ran until March of 2003 without ever going down. Not once. So it has a perfect operational record.

GAM: That's great! I remember during a visit withy Norman for something and he was having trouble finding the operating system. It had the feature when it sensed the machine was a little flaky it would take its tables and go to another machine.

LT: Well, yeah, the way that worked is any given instant it was controlled by a supervisor which ran at one point. But there were several other supervisors that were sleeping and the one that was in control would send sleeping pills to the others. So if anything terrible happened like the controlling supervisor went down or the network got partitioned because of lines going out then the other supervisors would wake up and contend for control of the network. One of them would win and the others would go back to sleep.

GAM: OK, that's a better explanation, yeah.

LT: So, whichever supervisor won it would take over control of the net. The changeover process usually took maybe two or three minutes. People who were on the network already during that time didn't see any disruption to the service but no new users could be accommodated until the takeover was complete. Also accounting information could not be acquired until the takeover was complete because the supervisor also coordinated the accounting information.

Time Net eventually got to about 5000 nodes and it was the time before the internet. The only other network at the time was the ARPA Network to, a lot of people have heard of the ARPA Network, not very many people today have heard of Time Net. But if you were a typical user of a terminal trying to access the mainframe somewhere you were almost certainly using Time Net, because the ARPANET was way too expensive and serviced mostly just universities. It did not service low speed terminals and it didn't service people who could not speak their particular language. Because Time Net was more than a communication system it was also a universal glue so that you could have a teletype, an ASCII Teletype and talk to an IBM 360 mainframe. Whereas the IBM 360 mainframe was expecting a 2741 with EBCDIC and you could take a 2741 with EBCDIC and talk to a UNIX system on a PDP-6 that knew only about ASCII Teletypes and Time Net would do all the conversions.

GAM: Now, to my imagination such programming problems are minutia deriven. That is they have a very broad horizon. You have to think about all these things, contingencies, and so forth. I'm impressed that you could keep all that stuff in your head. I don't see how to do it.

LT: Actually, the amount of code in Time Net was surprisingly small.

GAM: Maybe small, but it supported all these contingencies.

LT: It supported them well by design, rather than by the code itself.

GAM: Well, alright. It still seems like it's tough.

LT: Yeah, well in fact, a lot of the code, in fact, was conversion, conversion from this to that.

GAM: So, let me return to your career at the Lab. You were in a group that was being managed by Joe Wirshing, right?

LT: Joe Wirshing? We called him Bud Wirshing.

GAM: Yeah, I know.

LT: Yeah, Joe I guess is his proper name. My relationship with Wirshing was sort of strange. I think most people in the Computation's Division assume that mathematics and physics and that sort of thing was the Holy Grail that everybody wanted to be involved with, and machine shop support was sort of what the janitor did and it didn't have very much prestige. But somebody had to take care of the machine shops and the Metrology Department and all of that stuff, because otherwise the prototype widgets would never get built. So somebody had to do it and that fell into Wirshings domain and as long as I was able to keep everybody happy, Wirshing was happy. And as a result I never interacted with him very much. I could do pretty much whatever I wanted and the only time I met him, I might meet him in the hall passing by and he'd say hi and I'd say hi and that would be it. There was no such thing as an employee review or checking to see how well I was doing or any of that sort.

GAM: Did you ever interact with Sid Fernbach at all?

LT: Actually, the last time I saw Sid was at the CIA. And he was quite astonished to see me there. He wasn't expecting to see me there. He lost his composure for about half a second.

GAM: Was this at Fort Myer?

LT: Yes, No at the CIA Headquarters.

GAM: Oh, In Langley.

LT: In Langley, Virginia. But, I was there on a completely unrelated issue. But, other than that I think I only met him at a meeting once. At which meeting I said nothing and didn't really have anything to do with Sid, really.

LT: I'm sure Sid was very well aware of my existence, but there were two or three layers of management between me and Sid, and I really had no reason to interact with him directly.

GAM: Well, I understand, but you know Sid had been the head of Computation, such as it was, since November of 1952. He was around and actually a pretty good guy when it came to dealing with people. I know he liked to have special family gathering;, the Computation Department would have picnics and he was right in the thick of it. But at the Lab, he had an impossible job.

So, I'm running out of things that, I could challenge you this way I suppose. You summed up your experiences at the Laboratory and wanted to tell somebody about it, how would you describe that?

LT: Well, there are two completely separate phases and both of them very important to my personal growth.

The first phase was as a computer operator and the importance of that was, I guess I started out with an extreme inferiority complex, and had a very low opinion of myself. And when I got in the Lab I met people like Russ Peterson, who was in charge of the swing shift operators. He was a super guy. I met a lot of really great people there, Gary Tollefson, and several other operators and they kind of took me under their wing and entrusted these multi-million dollar machines to me and made me realize that I could handle it. And that was the first job that carried responsibility of that level that I'd ever had, and it was way beyond anything in my background that I could ever, where I came from in Michigan, I could ever even imagine such a thing.

GAM: Well, I don't know about that.

LT: It was a sort of breaking out of the mold that I'd grown up in.

And then the next step was after I graduated from Cal State at Hayward. The programming phase and that was where I got to learn all about the machine tools and the geometry and my operating system experience. I got just a touch of politics, not very much, but just enough to find out what politics were, certainly not enough to become expert at it. I got to travel as a programmer to places like Oak Ridge, mostly, but sometimes other places too. Their travel department was absolutely astonishing to me. There were, all the really gorgeous women at the Lab worked in the travel department. And when I had to go on a trip, they'd ask me where I was going and I would tell them where I was going and they'd say how much money do you want? And it never occurred to me that I'd even need the money, and I thought well, maybe I could get by with two hundred dollars and they'd say, "Oh, No, you need a lot more than that" and some absolutely stunningly gorgeous gal would open up her desk and there'd be a bunch of one hundred dollar bills there and she'd count a whole bunch of them and say, "Here, this'll probably take care of you" and I didn't have to sign for it or anything. And that kind of blew my mind. So I'd walk out with three thousand dollars of cash in my pocket.

GAM: Where in Michigan did you come from?

LT: It was a Dutch community, it was called Wyoming, Michigan, and it was a suburb of Grand Rapids, between Grand Rapids and Lake Michigan. Everybody I knew was Dutch, I was dimly aware that not all the people in the world were Dutch, but I wasn't quite sure of that. I never met anybody that wasn't Dutch until I left to go to the University of Michigan which was quite a shock and then when I came up to California that was even more of a shock.

GAM: The University of Michigan is forty thousand people.

LT: Actually, no, the University of Michigan is a lot more than that. Their stadium alone held a hundred thousand people. In fact, I remember standing in their stadium once at the beginning of a football game and I could see rivers of people flowing toward the stadium and I thought this is 0.6 percent of the population of the whole country, sitting in one spot. So anyway, I left Michigan in a 1963 Buick Special, brand new car, because I didn't enough money to buy a used car, so I had to buy a new car on credit.

GAM: You sort of had an idea that you were going to come to California, when you left Michigan.

LT: Yeah, but I hadn't heard of Livermore

GAM: Until you got out here.

LT: Yeah, I went for a job as a computer operator in Berkeley, Berkeley Lawrence Radiation Laboratory, and they told me about Livermore Lawrence Radiation Laboratory and that's how I found out about Livermore and drove out on highway 50 in my Buick Special. A two lane country road, highway 50 out in the middle of nowhere and they said, "Don't call us we'll call you" and that was the end of that, or so I thought. A few weeks later I got a letter in the mail saying, "Come out, we'll hire you" and because of that letter, I got to become a programmer.

GAM: I'm predicting that your story will charm the readers of our history site; I have the help of Tom DeBoni, Same Coleman, and my former secretary, SharonLee Danielson who will in fact, get this tape and produce a transcription that I will send it to you electronically for review. You may revise and otherwise edit as you see fit. After you return it to me, it will undergo semantic and syntactic checking and then it'll be sent to Sam will put it into HTML Tom will then post it on our site. I want to get a picture of you, I don't know if I have one or not so I have a camera here and we'll take a picture of you. I think we've run the gamut of things. I think there are a lot of interesting things, but I'm more or less constrained to take care of just Livermore.

LT: Yeah, I think we've covered the important part.

GAM: Yeah, well ok, I certainly appreciate you taking the time to do this; and finally, I want to thank you explicitly for this interview.




For information about this page, contact us at: comment@computer-history.info