Robert Richtmyer

Fitzpatrick:

Have to get started. What I typically like to do on these interviews, if you could start off a little by talking, for the record, about your background, particularly your educational and academic background in Physics, undergraduate and graduate studies.

Richtmyer:

Well, I was a graduate student in Physics at MIT from 1932 to '35. Got a Ph.D. in Physics, so I have M.A., B.A. — B.A., M.A., and Ph.D. in Physics.

Fitzpatrick:

All in Physics?

Richtmyer:

All in Physics. But part of my life I've been professor of Mathematics, and I'm not sure whether all of my colleagues always do have a degree in Physics.

Fitzpatrick:

Hmm. Were they all from MIT? All your degrees from MIT?

Richtmyer:

No. My first two were from Cornell. I was born and raised in Ithaca.

Fitzpatrick:

Oh, yeah. Mm hmm.

Richtmyer:

And then I was at Stanford in the Physics department, '35 to '40. And when the war came along I started working for the Navy department, in connection with magnetic and acoustic mine sweeping.

Fitzpatrick:

Going back to Stanford, were you at Stanford as a post-doc, or were you at Stanford as a faculty?

Richtmyer:

At Stanford I was in the faculty in the Physics department. So then, during the war I continued being a physicist.

Fitzpatrick:

Yeah.

Richtmyer:

And then I went to Los Alamos, at the end of the war, and in 1953 thereabouts, Koorot, Richard Koorot of New York University, gave me an appointment as associate professor in mathematics. So, at that time, I became a mathematician, and my first thought was, "If I'm going to be a mathematician, I've got to learn some mathematics." Now, as a theoretical physicist, of course, I knew a lot of mathematics, but I had never learned it quite the way mathematicians do it. And the NYU people were very kind to me. They let me teach the advanced courses, a lot of advanced courses in mathematics, and I had some good students who got, Ph.D. students who worked for me, and things of that sort. So that's my history.

Fitzpatrick:

Robert Richtmyer discusses how he ended up at Los Alamos and his friendship with Edward Teller.

Mm hmm. I'm curious, thought, to go back, how did you end up coming from Los Alamos? Who recruited you?

Richtmyer:

Let's see now. Well, I knew both Teller and Gamov, and they were both at George Washington University when I was in Washington, so I got together with them, you know, a lot of Physics, attended seminars there. And I think my going to Los Alamos was partly influenced by Gamov — I mean, by Teller, and Teller and I were quite good friends, then. And we continued to be good friends for some years, and then after a while things changed.

Fitzpatrick:

How's that?

Richtmyer:

Well, he just changed, you know.

Fitzpatrick:

When did he change? I'm not sure what you were —

Richtmyer:

Well, it was a gradual thing; partly when he turned against Oppenheimer, you now, all that sort of — So he and I just ceased having that kind of... So, I've never quite understood what happened there. But during the early years, he was a wonderful guy, you know? He took interest in young people and helped them and all that sort of thing, and he was a good physicist, and he played the piano, with his right foot rather heavily on the...

Fitzpatrick:

Because he has an artificial foot, yeah.

Richtmyer:

...on the pedal. But with great interest and enthusiasm.

Fitzpatrick:

So, it was mostly through him, you're saying, that you probably came to Los Alamos?

Richtmyer:

Well…

Fitzpatrick:

Not so much Gamow. I don't think Gamov was there at the end of the war, though.

Richtmyer:

No, no he wasn't. Not until later.

Fitzpatrick:

I thought it was.

Richtmyer:

And I had met Oppenheimer in my Stanford days.

Fitzpatrick:

Because he was at Berkeley.

Richtmyer:

At Berkeley.

Fitzpatrick:

Yeah.

Richtmyer:

And I used to go with some of my friends once a week, more or less, or every two weeks, to Berkeley to attend a theoretical Physics seminar there, when Oppenheimer was there, and he made a great impression on me. An incredible mind.

Fitzpatrick:

So, you come to Los Alamos around Spring of 1945, is that right?

Richtmyer:

Yes. No, between '44 and '45 I had made a few visits from Washington. Washington, I was associated first with the Navy department for several years, and then went OSRD. I worked with a group of people there, patent attorneys. Now what does 'patent' have to do with nuclear energy?

Fitzpatrick:

Well…

Richtmyer:

They were taking out guns; you know, the names of people like Oppenheimer, Teller, Gamov —

Fitzpatrick:

Von Neumann and Fuchs?

Richtmyer:

Von Neumann and so on.

Fitzpatrick:

[Laughs] So, you know, because according to David Hawkin's account, that Project Y history he did, he says you were hired into the patent office of Los Alamos.

Richtmyer:

Oh, that was when I was in Washington, D.C.

Fitzpatrick:

Okay. So you were more —

Richtmyer:

It was OSRD, it was a group of patent attorneys.

Fitzpatrick:

Right. So when do you move to T Division?

Richtmyer:

Well, then when I went to Los Alamos, for the first two or three months I continued working through the patent people with a man named Smith, Captain Smith, who was in charge of that. But then after a couple of months, I talked to Bradbury who was director or …

Fitzpatrick:

Well, Oppenheimer is head of this until about October of '45, and then Bradbury is appointed. I think October of '45 as a director.

Richtmyer:

And there was Bethe for example or something like that.

Fitzpatrick:

Yeah, uh huh.

Richtmyer:

And so after I'd been there about two months, three months maybe, I don't remember exactly, I took a position as an alternate T Division Lead. The leader at that time was —

Fitzpatrick:

Well, Bethe was leader during the war, but then George Plotchek —

Richtmyer:

Plotchek, yeah. Plotchek was the leader, and I was alternate, and then Plotchek left and I became division leader.

Fitzpatrick:

Well, that's what I was going to ask you, too. Plotchek was sick. I think he couldn't stand the altitude or something and had to leave.

Richtmyer:

Yeah, that sounds familiar. I didn't get to know him very well.

Fitzpatrick:

Mm hmm. Okay, see, but you were division leader only for about a year or so after the war.

Richtmyer:

Well, I think two or three years.

Fitzpatrick:

Two or three years? And then Carson Mark...?

Richtmyer:

Then Carson Mark took over, because I had decided to go to Princeton and work with Neumann and the thing that eventually was called 'hit ball'. And that was the big thing that I was involved in, by far the biggest thing. But I don't feel that I deserve all the credit.

Fitzpatrick:

I'd like to ask you more about that a little later, too. [laughs] By the end of '46, though, many of Los Alamos's staff had departed.

Richtmyer:

Oh yeah.

Fitzpatrick:

I mean, people were leaving, you know, they were just going to Cornell —

Richtmyer:

They were leaving [???], and I did a certain — I made several trips around the country, trying to recruit people.

Fitzpatrick:

Well, why did you decide to stay?

Richtmyer:

I'm not completely sure. But I've thought since that in a way it was a mistake, because see, the spirit at Los Alamos during the war was powerful, and there were refugees from Germany, and it was believed that the Germans were rather close to getting an atomic bomb, so therefore, it was goddamn, pardon my French, goddamn important to get there first! And people didn't work just 40 hours a week, you know. They worked day and night. And with great intensity, you know. And then of course, in 1945, all of that started to disappear, and particularly the people connected with employment and all that sort of thing came in. And during the war, they'd been very sort of informal and personal. Well, I guess there were laws, but then the security people came in and started making laws and all that, and the whole spirit changed.

Fitzpatrick:

It's gotten worse ever since, [laughs], so. Well, I was wondering if you could, you know, describe the general state, especially T Division, and during '46 and '47, because Carson has written that the number of full-time staff after '45 or '46 went from something like 35 to 8, and there was just, you know, no senior people, no senior theoreticians, and for this reason, he and others, and Bethe have claimed that that's the reason that worked just slowed right down to a complete stop. Although the fission and thermonuclear theory.

Richtmyer:

Yeah, well, I can't speak too much about the thermonuclear thing, but a lot of things continued and developed. Monte Carlo and the theory of ranching processes, which became a division of mathematics later, was started by David Hawkins and... I can't remember who.

Fitzpatrick:

Mm hmm. But after the war, that December '46, I believe you were also head of Team 9, is that right? You were Team 9, which was responsible for, quote-unquote, "advanced designs."

Richtmyer:

Responsible for what?

Fitzpatrick:

Advanced designs and weapons programs.

Richtmyer:

That doesn't ring any bell with me at all. I was —

Fitzpatrick:

Because Hawkins recorded this in his Project Y, he listed who was head of what particular group in T Division and other divisions.

Richtmyer:

I see.

Fitzpatrick:

And you were supposed to be head of Team 9.

Richtmyer:

Well, I was division leader of T Division, but I wasn't aware that I had any...

Fitzpatrick:

Maybe informally.

Richtmyer:

To further complicate — further connection with any group in T. Well, I knew a lot of the people, of course.

Fitzpatrick:

Mm hmm. Well, wasn't there also just, you know, right after the war there was a real lack of direction from Washington, and I mean, you know, the Manhattan District coming to an end, and then there's no big AEC officially until early 1947, and I think Bradberry is very worried about what's going to happen to Los Alamos. So I think initially Oppenheimer recommended to close it down right at the end of the war.

Richtmyer:

I'm not really aware of any of this. Of course, it's probably all in this book. Do you know this book?

Fitzpatrick:

Oh, yeah. Mm hmm.

Richtmyer:

Dark Song.

Fitzpatrick:

Right.

Richtmyer:

Not very well written.

Fitzpatrick:

It's hard to follow, and it's — He's got three different stories in there, and they —

Richtmyer:

Things are mixed up and he doesn't separate things to make them clear.

Fitzpatrick:

Mm hmm. I'd like to talk about that later a little bit, because there were some folks at Los Alamos, and you know, I think there's some critical parts in it as well, so. And his account of the Soviet program isn't really that good, so.

Richtmyer:

Yeah. Well, that I have no knowledge of, of course.

Fitzpatrick:

Maybe I'll move on a bit, then, to another topic I would like to talk about. If you may recall, one is the 1946 super conference.

Richtmyer:

Is the what?

Fitzpatrick:

The 1946 super conference, in April of 1946.

Richtmyer:

Yes, I remember that.

Fitzpatrick:

You attended.

Richtmyer:

Pardon?

Fitzpatrick:

You attended this.

Richtmyer:

Yes, I attended.

Fitzpatrick:

And a lot of Teller's group in T Division.

Richtmyer:

I remember it being quite interesting, but there was no violence, you know. Just a technical scientific meeting, really.

Fitzpatrick:

Mm hmm. Well, there's a couple questions I — Actually, this is the declassified version of the meeting LA-575. Unfortunately, it's been so much declassified and there's not a lot of substance to it, but there's a list of the attendees, at least, on the front.

Richtmyer:

Carl Metz, Norris Bradbury — there's something funny about that. Donald Flanders, Frankel...

Fitzpatrick:

Stanley Frankel.

Richtmyer:

Stanley Frankel, David Froman, Klaus Fulkes, Andy Horowitz, Dave Judd, Joe Keller. I knew all these people. Manley, Mark, Metropolis, Morrison. All these names are familiar. I knew them all. Moritiss, Serber, Teller, Tug, Turkenvitch and Moolah, Rinshirt. Stanley Frankel is listed again for some reason.

Fitzpatrick:

I think he's one of the authors of the report.

Richtmyer:

Oh.

Fitzpatrick:

It had several authors, although Teller apparently was the main author.

Richtmyer:

Tergevich and Nylar. And now its unclassified 1971.

Fitzpatrick:

Well, this version is unclassified.

Richtmyer:

This version.

Fitzpatrick:

If you look through it, there's just whole pages missing. I mean it's just; it's not very informative now. The original is still classified. Anything, nearly anything to do with thermonuclear, even if it's to do with the super-divisional stuff is still classified.

Richtmyer:

And I remember the — I have a copy of the shockwave manual, of Bethe and...what was the other? Lyle Idess [?], and it's been declassified. Before that, everybody knew it was C-classified, a well-known thing. It was on my desk, but the particular copy that I have in my desk hadn't been stamped properly. So I got into trouble. It's a black mark against me. [Laughs]

Fitzpatrick:

[Laughs] You're not the only one. Well, Teller has claimed that this conference, the April conference, was more Oppenheimer's idea, and others have told me that it was more Teller's idea, and I suspect the latter. Do you have any recollection of that?

Richtmyer:

No, I don't.

Fitzpatrick:

Mm hmm. Server, when I talked to Server, he said the same thing, and he thinks it was more because Teller didn't really get to work on super during the war, so...

Richtmyer:

Well, at that time, so far as I recall, Teller was on good terms with the other people involved.

Fitzpatrick:

Yeah.

Richtmyer:

And there were several people that were sort of leading the whole thing, and it was a nice, technical, scientific conference.

Fitzpatrick:

Well, the technical aspect of it, too, is interesting, because you know, to prepare for this conference, Metropolis, Stanley Frankel, and Anthony Turkevich go back to University of Pennsylvania, and they've run about eight calculations on the ENIAC [?].

Richtmyer:

Whose calculations?

Fitzpatrick:

They run about eight or so, I think, number.

Richtmyer:

Oh.

Fitzpatrick:

Eight or so calculations to see if the super will ignite using different parameters. And this has gone through Von Neumann, this is Von Neumann's idea, introduce them to the ENIAC at Penn. And they come back and present the results at this conference, ENIAC's results.

Richtmyer:

I don't recall that.

Fitzpatrick:

You don't recall that? And the general idea, if you read the classified version of this, is that it says the super will work using a few hundred grams of — ignite, using a few hundred grams of tritium, because tritium's a problem, of course.

Richtmyer:

Oh, yeah, sure, because it only lasts…

Fitzpatrick:

Mm hmm. And it makes the way the conference reads; it's a very optimistic conference. I mean, do you remember —

Richtmyer:

I remember us being —

Fitzpatrick:

— feeling that, yeah, recollection about it, uh huh. But I was wondering, too, if this is just, again, Teller, because when I interviewed Server last fall, he told me this story. He told me that after the super meeting, Teller wrote up a kind of rough-draft for this report. He sends it out to several people, including Server, to have them do corrections for it and to look it over. Server remembers that the thing sounded far too optimistic, saying that the thing will probably work, and he wrote to Teller saying you really out to change the report. And by the time it was completely edited, it was the same exact report that Teller had originally filed.

Richtmyer:

Are you in touch with Dave Hawkins?

Fitzpatrick:

I have not met him, no. I have not met him, and I probably should. [laughs]

Richtmyer:

No, I think, I assume he was present at the time of that conference.

Fitzpatrick:

He certainly knows —

Richtmyer:

His name is not on there, is it?

Fitzpatrick:

I don't know if he was at this meeting. Let's see if he was at this meeting.

Richtmyer:

So, what times —

Fitzpatrick:

No. But he knows a lot about the early super program. I mean, I'm aware of that, so I'm aware of that.

Richtmyer:

I think he knows a lot about everything, because he took the history seriously and wrote about it.

Fitzpatrick:

Right. And he wrote a classified history, too.

Richtmyer:

Yeah.

Fitzpatrick:

Part of that official history, and also for the Manhattan District history, the Army's history.

Richtmyer:

When was he writing that stuff?

Fitzpatrick:

'47, '48, you know.

Richtmyer:

Yeah.

Fitzpatrick:

Late '40s.

Richtmyer:

So he probably was in T Division at the time of that conference.

Fitzpatrick:

I don't know what division he was in, actually. I don't know if he was in T Division. I'm not too sure about that.

Richtmyer:

I'm not completely sure. I thought he was.

Fitzpatrick:

Mm hmm. Well, you do remember the meeting being kind of an optimistic one, so...

Richtmyer:

Yeah.

Fitzpatrick:

Yeah. And —

Richtmyer:

At least not pessimistic.

Fitzpatrick:

Right. All the others agree. Metropolis told me the same thing; others have said the same thing.

Richtmyer:

Have you been in touch with Nick?

Fitzpatrick:

Yeah. He still comes into his office every day or so. He's not too well, but he still comes to work, so I see him at lunch sometimes. I have lunch with him occasionally.

Richtmyer:

??? ??? ??? Roger Lanzaris [?] on the scene back then, at that time?

Fitzpatrick:

I don't think so. He showed — he told me he came in '50 or '51.

Richtmyer:

Ah hah.

Fitzpatrick:

He didn't show up until then.

Richtmyer:

Okay.

Fitzpatrick:

He came directly from Harvard.

Richtmyer:

Pardon?

Fitzpatrick:

I think he came directly from Harvard. He graduated from Harvard. But it wasn't till early fifties; I think he was around the forties.

Richtmyer:

So who else...? Hardly anybody left.

Fitzpatrick:

Yeah. I talked to Anthony Terkovich last year, too.

Richtmyer:

Oh.

Fitzpatrick:

He didn't remember too much, so...[laughs] But the conference's purpose, really, was it for kind of a feasibility study, not really a designed study.

Richtmyer:

Right, right. And discussion of some of the phenomena of theoretical things.

Fitzpatrick:

Which you didn't really know much about then, because according to Metropolis, and others, even when these problems were run on ENIAC, a lot of the affects have to be neglected.

Richtmyer:

Oh yeah, oh yeah.

Fitzpatrick:

Thermonuclear effects, [???], things like that you couldn't include. I mean, you would have overloaded the machine.

Richtmyer:

No, no no no. Back then they didn't know how to put things together.

Fitzpatrick:

Mm hmm. It's a very simplified; I mean technical aspect of it, at least as far as calculating —

Richtmyer:

In a way that was my big contribution. The phenomena, the simple phenomena involved three or four, about three, and for each one, the mathematical methods were known, partly, largely because of things done at Los Alamos.

Fitzpatrick:

Yeah.

Richtmyer:

Under-simplifying assumptions about the others. For example, if you have fissionable material, uniformly distributed, not moving, stationary, then you can compute the neutron multiplication.

Fitzpatrick:

That's reasonably easy.

Richtmyer:

So there were these things. The problem was to put them together for the big computers that were just then, and that's what I said. Well, anybody can say that. It just takes one sentence, and there must have been lots of people thinking about it. So I don't claim a lot of credit. But that was the big thing I did.

Fitzpatrick:

Mm hmm. Well I'd like to talk to you about that eventually, too, because that's important. One other thing is, according to this, I won't harp too long on the super report, but I just have a couple of questions about it, according to that report, the members of conference —

Richtmyer:

Members of?

Fitzpatrick:

Who was at the conference? All this group of people, the people at this conference, this 1946 conference, conclude that if you really wanted to build Teller's super it would take a year, perhaps two years, working at the pace that was done during war time for the fission program, and perhaps it might go. How serious was this taken, do you recall? I mean, how serious was this idea taken during the talk?

Richtmyer:

I can't really — but it could have been taken too seriously.

Fitzpatrick:

Mm hmm. Teller took it seriously. [laughs]

Richtmyer:

Nobody can predict how things are going to go.

Fitzpatrick:

Right, right. Well, we'll move on to the alarm clock, and see if you —

Richtmyer:

You know, my lack of memory is not just because with old age, memory disappears, but in addition I've lost interest in all those things. I haven't had any official contact with those people. I resigned my contract, I don't have any clearance of any kind anymore, I haven't given any thought to [???]. So that's in addition to old age.

Fitzpatrick:

Well, let me ask you this. Teller leaves for the University of Chicago in 1946. You know, Ferme, and Frankel goes there too, and probably a handful of other people. Teller was back in Los Alamos in 1993 giving a classified lecture on the origins of thermonuclear weapons.

Richtmyer:

'93?

Fitzpatrick:

He was back for the 50th anniversary.

Richtmyer:

Ah hah.

Fitzpatrick:

Mm hmm. Gave a talk. And Teller claims that after he leaves for Chicago, that Brichtmeyer is the person who holds the torch for thermonuclear weapons after the war, and Brichtmeyer actually did most of the work on the alarm clock. Is this true?

Richtmyer:

I think it's true. I think that has disappeared and I can't find it, because of my feeling that I acquired at some point, that it was an idea that just wasn't working. Wasn't worth following up.

Fitzpatrick:

Mm hmm, but in 1946 or '47 you have thought it would work. You didn't know.

Richtmyer:

Well, I did calculations on, I believe, on UNIVAC at New York University.

Fitzpatrick:

But this is before there's a UNIVAC or anything in existence. I mean, this is before even the SSEC is on the line, right at the end of the war.

Richtmyer:

The SSEC went on line, which was used for the Hippo Project.

Fitzpatrick:

Right.

Richtmyer:

That was the big thing, not just a few hours a day. Seven days a week, 12 hours a day, of 24 hours a day, people going in and out.

Fitzpatrick:

Right. But Teller claims that the first idea for the alarm clock comes to him in August of 1946.

Richtmyer:

August of 1946.

Fitzpatrick:

It's a year after the end of the war, which is pretty early.

Richtmyer:

Yeah, well —

Fitzpatrick:

I mean, he made the proposal only to you. It was between him and you.

Richtmyer:

Isn't that a [???].

Fitzpatrick:

That's what he claims.

Richtmyer:

Well, Teller and I were in touch at that time, there's no doubt about it. We were good friends then.

Fitzpatrick:

Well see, this original —

Richtmyer:

Can you tell me what the alarm clock is? I don't even know.

Fitzpatrick:

I have an idea, and I can give you a general idea. I'm supposed to be careful how much information I give I suppose, but anyway —

Richtmyer:

I don't talk.

Fitzpatrick:

The original — You can say whatever you want, because you're not an employee of the lab anymore, but the alarm clock, as I understand, in 1946, it's a very, very different configuration that what it involved into, which is the Texas -14.

Richtmyer:

Texas-14?

Fitzpatrick:

Which was the layering idea, with you and Douterai. But this is later. This appears later. But in 1946, the alarm clock report that I saw, I think it's called LA 6-10, it is also referred to as a 'Swiss cheese' configuration. And what, as I understand, it's supposed to be is almost kind of a fission bomb with just normal uranium, uranium 238, with pellets or lumps of deuterium placed within it. And supposedly, if this thing could be ignited, it would cause, you know, almost like a boosted bomb, I suppose.

Richtmyer:

Yeah.

Fitzpatrick:

It would help, you know, it would help create a detonation explosion more maybe on the order of a megaton, or megatons rather than kilotons, because of that presence of that deuterium.

Richtmyer:

Yeah, yeah.

Fitzpatrick:

And everybody wanted to use deuterium because it's easy to produce at this time. Tritium is not. And I'm wondering if this is proposed because it doesn't use tritium, as I understand. And it's an alternative to Teller's super, which is too expensive and —

Richtmyer:

Yeah, yeah. Anything that will avoid using tritium, anything that will avoid separating the U2-35 from the U2-38 had value, obviously.

Fitzpatrick:

Mm hmm. And I think maybe you did believe, at least on paper, you might have believed that it would work, because you wrote in this first report on the alarm clock in November of 1946, you wrote that this proposal was entirely feasible.

Richtmyer:

I wrote on the alarm?

Fitzpatrick:

And I quote, "Provided that there was not too much mixing of deuterium and the uranium by tailor [?] instability and light phenomena." Those are your words, unquote.

Richtmyer:

I don't remember that at all.

Fitzpatrick:

But at some point, at some point between then, between 1946 and probably '51 or '52, the name for this thing is kept, 'the alarm clock' is kept, and I've been told numerous times Teller named it this because he wanted to quote, "wake people up to the nuclear…"

Richtmyer:

Yeah, yeah, yeah, yeah.

Fitzpatrick:

You know, that's an old story. At some point, by '51 or '52, when you all are running these problems on CAC and UNIVAC, someone has come up with the idea of changing this configuration to have these layers, the onion-type of configuration.

Richtmyer:

And that was Teller's idea, wasn't it?

Fitzpatrick:

I don't know. That's what I was wondering, if you recall. And I've never seen a document that fully explains this or really gives credit to anybody, because at some point, someone comes up with the idea of using the theme deuteride, and you don't have to worry about this, you know, these pellets of deuterium mixed within uranium in layers. And again, layers of —

Richtmyer:

Spherical layers?

Fitzpatrick:

Mm hmm.

Richtmyer:

Or plain layers?

Fitzpatrick:

Some spherical layers.

Richtmyer:

Yeah, yeah.

Fitzpatrick:

Yeah.

Richtmyer:

Some calculations made them with plain layers.

Fitzpatrick:

That's possible, mm hmm. But do you remember who proposed using LID?

Richtmyer:

Lithium deuteride?

Fitzpatrick:

Lithium deuteride.

Richtmyer:

I have no idea.

Fitzpatrick:

Mm hmm. And some — you know, again, I'm trying to track down some time this configuration changed, and it becomes a TX-14.

Richtmyer:

You see, I feel I'm just failing you completely, because all this is gone.

Fitzpatrick:

No. I have lot's more questions on the CAC and on Hippo and all kinds of things like that, so don't worry. Don't worry. I'm just wondering if I can jog your memory, so... [Side B] All right, we'll skip ahead to about January of 1952. You, along with Roger Lazarus, Lotar Nordheim, several others, and you filed a report at Los Alamos on calculations that are under way on the standards user automatic computer in Washington, the CAC. And these are calculations on alarm clock systems.

Richtmyer:

We had many problems going on that machine.

Fitzpatrick:

Besides the alarm clock?

Richtmyer:

Yes.

Fitzpatrick:

Yeah.

Richtmyer:

And many different people were using that machine. I used to get there around 8:00 p.m., and I worked the machine until 8:00 a.m.

Fitzpatrick:

Yeah, Roger told me you'd use it all night. You were only allowed to use it at night.

Richtmyer:

And then I would go to a restaurant to have breakfast, and I'd order a martini. I'd get funny looks from the waitresses!

Fitzpatrick:

It was, yeah, it was downtown Washington this thing was located, right?

Richtmyer:

Yeah.

Fitzpatrick:

It was in downtown D.C.

Richtmyer:

Yeah.

Fitzpatrick:

And anyway, the first problems on the CAC, though, this group reports look exciting, and this is the first time, I think, that it's indicated that a lithium deuteride alarm clock may very well work.

Richtmyer:

I just have no memory.

Fitzpatrick:

But it's interesting, because there's this configuration. By this time, in early '52, the preparations for the Mike shop, the giant deuterium fuel device that's going to be shot in November on Bikini, the one that employs radiation implosion, this stuff is all under way, of course. And this thing, you know, this thing was giant, this huge cryogenic assembly. I don't remember how many tons this weighed. It was just undeliverable; you couldn't deliver it by any aircraft at the time, right? It's this big operation. But how come — what I'm wondering is, how come an alarm clock isn't tested first, if this idea is so far along, apparently? Why do they bother to go ahead testing Mike? I don't know if they wanted proof of principle or...

Richtmyer:

I didn't want to guess, but I think there might have been some feeling that the Mike shop, eventually would lead to things that could be traced, boarded by aircraft.

Fitzpatrick:

Right. Well, and apparently there was a program to try and weaponize Mike, following its test.

Richtmyer:

I wasn't there.

Fitzpatrick:

Well, I think Jay Wexler and some of those —

Richtmyer:

Clearly not something you would carry by plane.

Fitzpatrick:

No. [laughs] No. But you know, the alarm clock was, I think, in principle.

Richtmyer:

Yeah, in principle.

Fitzpatrick:

But you know. But it's tested later.

Richtmyer:

Why it wasn't tested, I don't know.

Fitzpatrick:

At least not this early, it wasn't. I was wondering why it wasn't tested first. And then you described it, I think in that report on LA 13-57, on the CAC calculations, your report, and the others' reports, it's a rugged, and you say it's easy to build.

Richtmyer:

On which?

Fitzpatrick:

This is this report, this —

Richtmyer:

On the CAC.

Fitzpatrick:

The CAC calculations, mmm hmm. Because apparently it's already quite advanced, the alarm clock's already quite advanced, even before radiation implosion has been clarified, or put forth. Which we still don't know who proposed exactly what or who was totally responsible for that.

Richtmyer:

I refuse to take part.

Fitzpatrick:

Yeah, I understand.

Richtmyer:

Take sides.

Fitzpatrick:

Uh huh. That's okay. But maybe more directly, how did you get to use a CAC? How did you get —? Did you know someone there, or was it through Von Neumann, or...?

Richtmyer:

I don't remember, but there was a big group of people who sort of all knew each other and we knew about it, and in particular, there was one of the engineers at the institute project in Princeton.

Fitzpatrick:

The IAS machine, you mean.

Richtmyer:

Um hm, who were living there, and moved to Washington to work on the CAC. And he finished the CAC before the machine in Princeton —

Fitzpatrick:

Oh, the Princeton machine was just — That was a funny story, too, because that—

Richtmyer:

And that engineer was here at the Bureau of Standards now. I can't think of him name right off.

Fitzpatrick:

That's curious, because you know, that's interesting too, because Von Neumann had proposed to build the IAS computer in 1945, I think.

Richtmyer:

Oh yeah.

Fitzpatrick:

And he got money from ONR and —

Richtmyer:

And all the basic ideas that went into computers for a long many years, were his ideas.

Fitzpatrick:

Sure. I mean, MANIAC [?] was modeled on that.

Richtmyer:

Oh yes, definitely.

Fitzpatrick:

And the JOHNIAC [?] and all of those several machines built right after. But funny, I mean, MANIAC and others were finished before Von Neumann finishes the machine at IAS.

Richtmyer:

Yeah. Well, Von Neumann was not the engineer. The engineer —

Fitzpatrick:

No, there was another guy that was working on that as well, and then, you know, I'm sure several engineers too, but...

Richtmyer:

Many engineers, yeah, and there was... Well.

Fitzpatrick:

So anyway, Roger Lazarus tells me, when I talked to him, told me several things about this program you're writing on CAC in '51 and '52, on the alarm clock, or one of the problems is the alarm clock, and you could only run your calculations at night.

Richtmyer:

Well, that's because somebody else has the machine during the day.

Fitzpatrick:

Roger told me you also had problems with the CAC, and I quote from him. He said, "It was a four-address machine, and A and B is argument, C is your result, and D is where to take the next command."

Richtmyer:

Yeah.

Fitzpatrick:

Thus, he says, "you can jump around." And he tells me, "I managed to get some of the instructions to have a numerical value of some of the constants, so we could double up the constants with the instructions and fit it into 10-24. We worked for a month with only one good run," he says, and that run was just long enough to let the experts back here deduce there was a factor of ten error in one constant. "Error was equivalent to having the wrong speed of light," he says. So the explosion front was moving too slowly, he says. "We never could get the CAC to go again." Is that true?

Richtmyer:

I don't remember that.

Fitzpatrick:

Because I thought you had some success with that machine, you know, and I...

Richtmyer:

In a way the machine itself didn't interest me too much, because I knew all Von Neumann's ideas and the plans.

Fitzpatrick:

Right, right.

Richtmyer:

And the CAC was a temporary sort of thing, in a way.

Fitzpatrick:

What do you mean by that?

Richtmyer:

Well, it wasn't the machine that the Von Neumann people were planning.

Fitzpatrick:

Right, right.

Richtmyer:

It was only a particular; the input and output were run on teletype machines. You see, my — the big thing that I did was the Hippo thing. That development was — well, again, I can't claim all the credit. But that was a big thing, and the methods that were developed there were used subsequently in Los Alamos and other places by the Atomic Energy Commission for quite a few years.

Fitzpatrick:

Let's skip to that topic then, and we can talk about Hippo. Just let me skip back in time. Did you alone first propose to do the Hippo program?

Richtmyer:

Well, in a way I did, yes. I proposed it to Von Neumann. Well, I came along at the right place at the right time, so to say, and all of the ingredients were there. The problem was to put them together, and that would take a bigger computer than anything that existed then, and that was what I proposed. And as I say, other people must have had the same idea, but somehow I was the one that proposed it to Von Neumann, and he liked the idea.

Fitzpatrick:

And Von Neumann, though, I think Von Neumann originally wanted to run that on ENIAC.

Richtmyer:

We talked about running it on the ENIAC, right. During the first — See, I was there in Princeton for three years, I think. In the beginning we had in mind the ENIAC.

Fitzpatrick:

Now, when did you go to Princeton?

Richtmyer:

'47.

Fitzpatrick:

And you were visiting. You were back and forth between —

Richtmyer:

Well, I officially was started as a visitor to the institute to work with Von Neumann in '47, and then that work continued partly at Princeton, partly in New York City, when we went to the IBM machine. But I was working in a sense with Von Neumann all that time. Incredible man. Incredible!

Fitzpatrick:

That's what everybody's told me. I know. I wish he were still alive. It would be wonderful to interview him. But Hippo, I assume Hippo gets its name because of its large size.

Richtmyer:

Something I gave you here...

Fitzpatrick:

I think this — I put it over there.

Richtmyer:

Oh. Origin of the name Hippo.

Fitzpatrick:

Okay.

Richtmyer:

You can read it faster than I can tell you.

Fitzpatrick:

Okay, let's see. In my work for Los Alamos at Princeton in 1947 to 1950, I was joined by Adell Goldstein and Clara Von Neumann, who shared my office with me. I had a habit of writing on the upper right corner of the blackboard cryptic notes to myself about things I had to do. On one occasion I was away for about ten days, and when I returned, there was an additional note on the blackboard, an imitation of my handwriting, said, Fresh water for Hippo. In consequence, Hippo became the code name for the project we were working on.

Richtmyer:

Yeah. Now, I'm not sure whether that —

Fitzpatrick:

So it might have been one of those ladies who proposed the name.

Richtmyer:

Either Clara Von Neumann or Adell Goldstein, or between them, observed that. That was their doing. It was just a joke, but then that became the official name.

Fitzpatrick:

Yeah. Well there was also, though, there was a baby Hippo project, too. There was a baby Hippo that Walt Lanshoff [?], who stays back at Los Alamos, proposes.

Richtmyer:

Ah hah.

Fitzpatrick:

Do you know about that?

Richtmyer:

No I don't. But what machine did he have in mind?

Fitzpatrick:

I think what that was for, while, you were back in Princeton working on Hippo, Lanshoff proposes this thing that could be run on IBM punch card machines only.

Richtmyer:

Yeah, well, Los Alamos had IBM punch card machines, of course, at that time. They used them for the implosion work.

Fitzpatrick:

Got them during the war, right.

Richtmyer:

I wasn't aware as Walt Lanshoff, but I don't see how there can be a baby Hippo. I mean, there were those three major phenomena, and the whole thing — They had to be run simultaneously, taking into account complicated interactions that take one or the other.

Fitzpatrick:

Now go back, three major phenomena. You're referring to a simulation of the trinity device. Isn't this what Hippo was for?

Richtmyer:

Well, yeah, well, they had in mind the trinity device, or any other fission bomb.

Fitzpatrick:

Okay. Well, right, the Hippo subtitle, I think, is called mechanized calculations of efficiencies, and other features of fission devices. I think.

Richtmyer:

Oh, well, I don't know what the official name was. But that was a big idea at that time. I say it was my idea, but there must have been other people with the same idea.

Fitzpatrick:

Well, sure, I mean, you have other people working on this with you.

Richtmyer:

And there were many, many other people working on it.

Fitzpatrick:

Foster Evans, Certa Evans.

Richtmyer:

Yeah, that's right.

Fitzpatrick:

Adell Goldstein, Herman Goldstein, Von Neumann.

Richtmyer:

Adell Goldstein and Clara Von Neumann were working with me under Johnny, and the four of us were — But then we had lots of other people, IBM employees. Now, that machine, SSEC, although primitive in a way —

Fitzpatrick:

It was excruciatingly slow, is that correct?

Richtmyer:

Oh yeah.

Fitzpatrick:

Yeah.

Richtmyer:

But it was probably the physically biggest machine ever built. It occupied three floors.

Fitzpatrick:

Didn't it sit on Madison Avenue as kind of a show-piece, too? Wasn't there a glass rung to it or something?

Richtmyer:

Well yeah, well that's right. There was a show-piece aspect of it, and it had 40,000 vacuum tubes. And they weren't miniature tubes, either.

Fitzpatrick:

They were the large Williams [?] ones.

Richtmyer:

And 10,000 relays and — It was physically the largest machine ever built, and we used it 24 hours a day for a couple of years.

Fitzpatrick:

Right. How did you get to use SSEC? Did someone have to go to Tom Watson, or did...

Richtmyer:

I don't know what was the official —

Fitzpatrick:

Von Neumann might have had that connection?

Richtmyer:

But the AEC arranged it somehow.

Fitzpatrick:

The AEC arranged it.

Richtmyer:

And the IBM employees, they were not cleared on the classified aspect. They didn't know what we were doing. I'm sure some of them guessed.

Fitzpatrick:

Yeah, but originally, when this group went to Princeton to work with Von Neumann, was the original intention to use Von Neumann's computer with the idea that it would be completed? And then you have to go over to SSEC?

Richtmyer:

At first, there was a big mathematical programming. It was somewhat independent of which machine you use.

Fitzpatrick:

I see.

Richtmyer:

But I think we had in mind the ENIAC at that time.

Fitzpatrick:

I believe you did, yeah.

Richtmyer:

Or the Princeton machine, if it should get finished.

Fitzpatrick:

But the ENIAC couldn't handle this program.

Richtmyer:

The ENIAC was too small, much too small.

Fitzpatrick:

Right.

Richtmyer:

ENIAC was a good machine.

Fitzpatrick:

Yeah. Well, and interestingly, Los Alamos is using this thing from 1945 up through '49 or '50, and I'm interested in that, all kinds of things, I mean, lots of folks are going back in with Foster Evans and Certa Evans. Metropolis goes back a few times.

Richtmyer:

Yeah, yeah. That's right.

Fitzpatrick:

They're running implosion calculations, they're running hydroid calculations.

Richtmyer:

Yeah.

Fitzpatrick:

They're running all kinds of things back here, and I was wondering, why — I mean, could this be because there's not enough capacity on the punch card machines at Los Alamos, or they just, you know, there's such an interest in computing now, people want to use this thing.

Richtmyer:

Yeah. Well, thing is, when you put those three phenomena together, each one is non-linear because of the interaction with the others and all that sort of thing.

Fitzpatrick:

The three phenomena you're referring to, specifically?

Richtmyer:

Neutron diffusion…

Fitzpatrick:

Neutron diffusion, uh huh.

Richtmyer:

The energy flow within the thing, and projection of shock waves and that's three. I don't know what they're counting, but if you put all those things together and take them seriously, you have an enormous thing. And at that point, we were concerned mainly with the programming aspects of it.

Fitzpatrick:

Right.

Richtmyer:

Whether it could be done on the ENIAC wasn't completely clear at that time.

Fitzpatrick:

Right.

Richtmyer:

But then somehow we switched and decided to put it on the SSEC, and I'm sure Johnny Von Neumann was involved in that part of the —

Fitzpatrick:

He knew everybody, yeah.

Richtmyer:

That's right.

Fitzpatrick:

He, as I understand, he had a lot of connections, you know, to NBS, to IBM.

Richtmyer:

That's right.

Fitzpatrick:

He'd been a consultant to the army, of course, during the war. That's how he got involved with ENIAC anyway, you know, with knowing Goldstein and Eckerd [?] and Malkley [?].

Richtmyer:

Yeah. And the ENIAC was a very interesting machine from many points of view.

Fitzpatrick:

Mm hmm. I've seen pieces of it, as a matter of fact. I was back at Penn last spring for — in fact, it was ENIAC's 50th birthday party.

Richtmyer:

Oh really.

Fitzpatrick:

And they have part of it reconstructed.

Richtmyer:

And Herman Goldstein and Adell, his wife, they both knew that machine very well. They worked with it.

Fitzpatrick:

Yeah, yeah. Hmm. So Hippo is supposed to give a clear indication of what's actually going on in a fission device, because —

Richtmyer:

Yeah, and —

Fitzpatrick:

All they had really was a kind of guess, you know, from the war time.

Richtmyer:

My feeling is that the word 'Hippo' should refer to a whole bunch of mathematical things.

Fitzpatrick:

I see.

Richtmyer:

Aspects, rather than any particular calculation or result.

Fitzpatrick:

Another thing, though, Carson has a short history of thermonuclear weapons that he's written.

Richtmyer:

Oh.

Fitzpatrick:

Carson wrote a very short history of the early thermonuclear project at Los Alamos.

Richtmyer:

Well, Carson was very good. He was very much in touch with everything that was going on in the T Division. An incredible memory for that sort of thing.

Fitzpatrick:

Master of the English language, too.

Richtmyer:

Yeah, yeah.

Fitzpatrick:

I've read a lot of his letters. But he suggested Hippo is also done — the second reason in mind, and he says that, you know, at this time, any initiator for a super, for a thermonuclear, the idea is to have a fission initiator, and you had to know exactly what was going on in that fission device…

Richtmyer:

Yeah.

Fitzpatrick:

…if you're going to ignite —

Richtmyer:

And that was the sort of thing that Hippo was designed for was to handle all those problems of that kind.

Fitzpatrick:

I also ask this because in the final Hippo report there's a description of actually two final calculations that are done. One is for a simulation of a trinity-type device, right? Another is for what is called a large, gun-type bomb. This is very vague, but it's known as a multi-crypt gun, also known as Project Little Edward. Was this also a main-proposed initiator for a super? Do you recall the calculation, too?

Richtmyer:

I don't —

Fitzpatrick:

The multi-crypt gun, or Little Edward?

Richtmyer:

But once the whole great machinery of the Hippo mathematical structure was known, it could be used for lots of things. Anything involving Hippo, a nuclear weapon.

Fitzpatrick:

Carson also suggests that in about 1948, it must have been when you just started working on this thing, on Hippo, you along with Von Neumann proposed, or introduced, a new technique in weapons calculations called artificial viscosity or shocks.

Richtmyer:

Oh yeah, yeah.

Fitzpatrick:

Which I have, actually, your reports on that have declassified. One of the few things that is.

Richtmyer:

It was a published paper, though.

Fitzpatrick:

Right. Here's part one, and this is part two.

Richtmyer:

Well, it was actually a published —

Fitzpatrick:

Yeah, these are just the LA reports.

Richtmyer:

Oh.

Fitzpatrick:

These are what are in our library. You may have to flip the page. They x-out so much stuff, it's impossible to read.

Richtmyer:

Oh yeah. That shock thing. That was used all over the world.

Fitzpatrick:

And this is yours, yours and Von Neumann's work?

Richtmyer:

Well, it was actually my idea.

Fitzpatrick:

Your work, uh huh.

Richtmyer:

But it came from things that I had been discussing Von Neumann.

Fitzpatrick:

Right.

Richtmyer:

And so we wrote the paper together, the first paper.

Fitzpatrick:

And this — but Carson suggests that you did this in connection with problems arising with Hippo. Is that —? What was the same opinion?

Richtmyer:

Yeah, I think we did. I think that came out just as one little detail.

Fitzpatrick:

But, I mean, can you elaborate on this anymore, or is this something — how this paper came about? Or do you recall?

Richtmyer:

Well, it was an idea that Von Neumann and I had between us, and maybe with other people. I don't know. And we sort of crystallized it and wrote about it, and then the paper we published.

Fitzpatrick:

Because I believe it was something that couldn't be handled on the punch card machines at all. That was — I think Carson talks about that.

Richtmyer:

To follow a motion of a shock involves quite an elaborate mathematical sort of thing. And our approach to it made it easier, that's all. We became sort of famous for that, in fact, more than I really wanted. But it was a good paper. But then other people wrote other versions of it, in particular...oh dear…

Fitzpatrick:

All right. Well, we can move on and talk about a little more about computing or computers in general, if that's okay.

Richtmyer:

You know, we made tea, but we didn't do anything.

Fitzpatrick:

We can have it now. We can stop. That's always fine. [Tape cut off]

Richtmyer:

...designed according to the trinity device.

Fitzpatrick:

But there was never another Hippo, I mean, the same sort of programming done again, is that correct?

Richtmyer:

I beg your pardon?

Fitzpatrick:

There was never another large Hippo program the same thing was ever done again, I don't believe. I mean, is this the only time that this is done?

Richtmyer:

Well, we did several programs, several calculations in New York City, at IBM world headquarters.

Fitzpatrick:

Do you recall what else you did?

Richtmyer:

I think we did about three, three calculations.

Fitzpatrick:

The one on the multi-crypt gun as well, I know that one, but I didn't know there was another one.

Richtmyer:

But I think of the word 'Hippo' does not apply any one of those calculations, but to the general thing that led to the methodology that had been developed.

Fitzpatrick:

Well a lot of these things were a problem. I mean, from the war time. I mean a lot of these calculations, because early on, even before Los Alamos opens, Eldred Nelson and Stanley Frankel are at the Berkeley conference, and they're trying to calculate neutron diffusion and are having a lot of trouble doing it on these desk calculators. And there's no way to accurately model this, you know, apparently.

Richtmyer:

Yeah. Well, if the material is of uniform distribution, and it sits still, doesn't move, then a calculation — that was not easy in the earlier days, but that exists and could be done.

Fitzpatrick:

Well, this is the reason, I think the reason, which punch card machines are ordered during the war.

Richtmyer:

Well, at Los Alamos, the punch card machines were used for the implosion.

Fitzpatrick:

Right.

Richtmyer:

And I think I told you that —

Fitzpatrick:

But I think they were ordered with neutron diffusion calculations in mind, because the implosion weapon wasn't taken seriously till 1944, because they were trying to build a gun device, a plutonium gun device at first.

Richtmyer:

And the gun device was never tested until Hiroshima.

Fitzpatrick:

Right.

Richtmyer:

But that's a very much simpler sort of thing. Once you have enough U2 35, something that could be built in any good machine shop.

Fitzpatrick:

But the original gun device that was talked about at the Berkeley conference in <42, with Oppenheimer and Teller, Von Flex, Server, a few others are there, they wanted to actually build a plutonium gun device, and then they figure out that plutonium is spontaneously fissioning, and the thing would just, you know, be a dud, so...

Richtmyer:

Yeah.

Fitzpatrick:

So that's when implosion, the program was switched to implosion and the punch card machines come in. Maybe, getting back to computers again, now when do you start working on the UNIVAC?

Richtmyer:

Well, I just don't remember. Oh, wait a minute, yes, yes. At some earlier point in all that I started working on the UNIVAC at the companies in Philadelphia.

Fitzpatrick:

That was Eckerd and Malfey's company they started.

Richtmyer:

Yeah. What was it called?

Fitzpatrick:

Birmington Rend [?].

Richtmyer:

Birmington — yeah. And we actually took over the machine they had out there, the machine there for showing off to people and all that, together with one or two of their employees. And then either that same machine or another one just like it was established at New York University.

Fitzpatrick:

Yeah, there was another UNIVAC put at NYU.

Richtmyer:

Twenty-five Waverley Place, New York City. And we used that for many, many years as well, I guess many years.

Fitzpatrick:

But by this time, you've left Los Alamos. Is that when you go to NYU, or are you still consulting?

Richtmyer:

Well, I was working for Los Alamos and various other places. But I didn't actually move to NYU. I didn't officially move to NYU till about '53 I think. But that UNIVAC was established in a building at 25 Waverley Place.

Fitzpatrick:

What's interesting, too, is that by the end of the war, the T Division itself, at Los Alamos is becoming interested in acquiring its own computer. You know, they're becoming interested in all these large machines in general, and the ENIACs at Penn, Von Neumann's machine, is supposedly under way at IAS and others. But some-time very soon after the war is over, there is all this talk of building a computer at Los Alamos, and in fact, I have evidence that Bradbury wrote to that AEC in 1946 saying that there was a loose plan for constructing a computer at Los Alamos. Do you have any idea who proposed to build a computer at Los Alamos, initially?

Richtmyer:

No I don't, because during the war that idea played no role at all. It was quite [???].

Fitzpatrick:

Well, this is after the war, though. This is —

Richtmyer:

After the war, of course, there existed the ENIAC and the plans for the Princeton machine.

Fitzpatrick:

Yeah. You know, and everybody's kind of connected in this.

Richtmyer:

And by then there was the IBM punch card. What was that method called? You could use…oh dear, what was it called? Computing system based on the IBM machines, and I actually used it for some purposes.

Fitzpatrick:

Computing system for...

Richtmyer:

You essentially have a punch card machine sitting in front of you on a desk, maybe some storage things, and they had developed a computing system. What do you call it? A computer program.

Fitzpatrick:

You don't mean the way with Feinman and his people...

Richtmyer:

...who was involved, I don't know.

Fitzpatrick:

Well, several programs, I think, were, you know, at some point after the war came of them were developed for the IBM machines, and then switched over and used on IBM's own electron computers, the 701 and 704.

Richtmyer:

Yeah, well, 701 and 704 were — Well, 701 was designed a look like an imitation of Von Neumann's ideas.

Fitzpatrick:

That's what I, yeah, that's what I was told by several people, so he influenced it very much. It was very similar to it. And in fact, I heard this story. I don't know if this is true or not, I asked Metropolis and he says not. But that came about 1950, '51 or so, when Metropolis and other folks are trying to complete the MANIAC, the Los Alamos MANIAC, IBM is getting very worried, because they're also at the same time developing their 701, and they want very badly to be able to rent one to Los Alamos.

Richtmyer:

And they eventually did.

Fitzpatrick:

And they did, but they were afraid of Metropolis, because if Jim Richardson got the MANIAC finished then they may not be able to rent one.

Richtmyer:

Yeah, but I don't think the Los Alamos MANIAC ever got, really, very much off that.

Fitzpatrick:

Well, you know, well — Yeah, you know, Roger Lazarus suggested that too, and Ed Borghese [?], if you remember Ed Borghese, he suggested that as well.

Richtmyer:

Yeah.

Fitzpatrick:

And I haven't really looked into that as much, but I've been told that some of the calculations for Mike were run on MANIAC, but after that it was really more of a recreational-type of computer [laughs] rather than a working computer. People were playing chess on it and such. You know, there are a lot of photographs of that as a matter of fact.

Richtmyer:

Yeah, yeah.

Fitzpatrick:

And you know, and George Gamov and Rena Gardner are those who're doing DNA calculations, and at the time DNA is discovered, and all kinds of intellectual pursuits, I guess. But as far as weapons calculations, I'm not too sure either.

Richtmyer:

And then the 701 came in, and that was a serious thing.

Fitzpatrick:

Oh, right. Yeah. But I have a letter. I have another memo that I dug up of yours that you wrote to Bradbury in about 1947, saying that you thought the planning of the computing lab at Los Alamos was very pertinent to Los Alamos's problems. So apparently you yourself were encouraging Bradbury to purchase, or you know, sanction construction of such a computer.

Richtmyer:

Yeah.

Fitzpatrick:

I assume you got interested in this through Von Neumann, or you know.

Richtmyer:

Oh yeah. Everything I did with him was through Von Neumann, really that way.

Fitzpatrick:

And Teller. I mean, Teller, too, is writing letters to Bradbury saying we really ought to get a machine here at Los Alamos. I assume this I is so — the idea behind it is so you don't have to farm calculations out to all these other facilities.

Richtmyer:

Yeah.

Fitzpatrick:

Because you're doing this —

Richtmyer:

And if it's right in your own place, you have more control somehow, and...

Fitzpatrick:

Right, right. But the MANIAC itself, I know the Los Alamos MANIAC is also based on Von Neumann's plan for the IAS.

Richtmyer:

Oh yes, yes. Essentially the same, I think.

Fitzpatrick:

So why does Carson bring Metropolis back to Los Alamos to lead this project? Do you have any idea?

Richtmyer:

To lead what project?

Fitzpatrick:

To lead the MANIAC project.

Richtmyer:

I wasn't aware that — What was the MANIAC project at Los Alamos?

Fitzpatrick:

That was the computer that was the mathematical and numerical integrator and computer. He names it MANIAC, based on being just like Von Neumann's machine at IAS, in order to do weapons calculations on the site of Los Alamos.

Richtmyer:

Was it a matter of transplanting what had been done at Princeton to Los Alamos, so to have a copy right there? Is that it?

Fitzpatrick:

I guess so. That's, you know, the early history isn't too clear to me.

Richtmyer:

I'm not aware that Los Alamos put any ideas into computer design that went beyond what Von Neumann would have.

Fitzpatrick:

I think they changed it. I think they actually did change. By the time the MANIAC was finished, there was some changes, and it wasn't exactly like the IAS machine. The size of the tubes —

Richtmyer:

I wasn't aware of differences there.

Fitzpatrick:

But that was, I think those were all Jim Richardson, the engineer Jim Richardson's ideas. But those are details.

Richtmyer:

Yeah, Jim Richardson was the important person there.

Fitzpatrick:

Mm hmm. He just died…

Richtmyer:

Just recently.

Fitzpatrick:

…like fall, yeah. There was an obituary in the IEEE in the History of Computing, which I don't know if you read. Okay. One other thing Carson also told me is that early Von Neumann, apparently Von Neumann kept promising to T Division that the IAS machine was going to be finished very early. For some reason he thought so, and everyone was anxious about that. Of course, it's not finished until, I don't know, late '52 or early '53.

Richtmyer:

And other machines were finished already before that.

Fitzpatrick:

Right, right, including the Los Alamos MANIAC and lots of others. But, now did you use Von Neumann's, did you use the IAS machine when it was done?

Richtmyer:

At Princeton? No.

Fitzpatrick:

You never did?

Richtmyer:

Never did.

Fitzpatrick:

Okay.

Richtmyer:

I was in the same building. I had an office right in that building where it was all going on, but the machine was never finished during my period.

Fitzpatrick:

I see. And there was no need to go back to use it once it was done, I guess, if you know, there's other machines available.

Richtmyer:

Other machines available.

Fitzpatrick:

Like UNIVAC, mm hm.

Richtmyer:

And little point there, this distinction between machine language and programming language. So up until a certain point, when you worked with a machine, you worked in the machine language. Nobody ever thought of anything else.

Fitzpatrick:

There was no such thing as software back then!

Richtmyer:

And in particular, the UNIVAC, the machine language of the UNIVAC, amazing flexibility, it was a marvelous thing to program in that language, very easily. For example, it would handle sub routines better than FORTRAN. It was a marvelous thing. But then when programming languages came in, then machine languages began to deteriorate, because they were used only for that. So I think the UNIVAC machine language was the peak of machine languages. Marvelous thing.

Fitzpatrick:

UNIVAC didn't last, though. I mean, as a lot of these, you know, they…

Richtmyer:

No. The UNIVAC at NYU, they developed a programming language for the UNIVAC.

Fitzpatrick:

Which was called?

Richtmyer:

Oh, I forget what it was called. But it was essentially just a mechanism for handling addresses and things like that. Might grow since then, but my God, FORTRAN lasted so long. I used it so many times, for example, either at NCAR, National Center for Abstract ???, when their big— What was that machine called? Ten years ago, so their machine was one of the biggest and fastest in the whole world. And how do people use it? Mostly by FORTRAN, and I programmed many problems in FORTRAN. And I hated it.

Fitzpatrick:

Mm hmm. I hated it to, when I had to. They were still using it when I was in college, you know, ten years ago, so we had to learn it. So it's been around a while. And what else?

Richtmyer:

Incidentally — Oh, it's only 10:30. Are you going to have lunch with me?

Fitzpatrick:

I'd love to.

Richtmyer:

Okay, great.

Fitzpatrick:

That'll be fine.

Richtmyer:

I don't know just what to fix, but will you help me in the kitchen?

Fitzpatrick:

Sure.

Richtmyer:

Sure, okay.

Fitzpatrick:

Sure, it's no problem.

Richtmyer:

We'll fix something.

Fitzpatrick:

Okay. One thing, though, I mean, is there — One of the justifications for using, for Los Alamos's using ENIAC, CAC, SSEC, there's UNIVAC, is that also to do with the staff shortage at Los Alamos? Because you know, post-war, there's just such a staff shortage. There are just a few people.

Richtmyer:

Well, I'm not sure I know how to answer that, except that during the war, and maybe for the next year, people were so busy with the Physics of the things they were doing that they didn't have time to think about designing computers. That wasn't their job, some of them.

Fitzpatrick:

Yeah, sure, yeah.

Richtmyer:

That was the job of the people at Aberdeen and Princeton and NYU and Chicago, so on.

Fitzpatrick:

Well, Von Neumann was doing both.

Richtmyer:

Oh yeah, sure.

Fitzpatrick:

He's designing computers and he's regular consultant to the laboratory.

Richtmyer:

Yeah. Well, when he was in Los Alamos, he functioned as a physicist.

Fitzpatrick:

Right.

Richtmyer:

A damn good physicist. One of the best. He knew everything in Physics. Incredible man. So he wasn't designing computers.

Fitzpatrick:

He got interested in that again sometime towards the end of the war, around when he met Goldstein.

Richtmyer:

Yeah.

Fitzpatrick:

And when he got involved in ENIAC, I think.

Richtmyer:

But that was not the sort of thing he was thinking about doing when he was at Los Alamos.

Fitzpatrick:

Well, he was helping design the lenses for the implosion device from [???].

Richtmyer:

Oh yeah. Right.

Fitzpatrick:

And several other things. But let's see what else I have.

Richtmyer:

Did you ever meet him?

Fitzpatrick:

Oh no. I'm afraid not. I think he died in 1957, '58, brain tumor.

Richtmyer:

That sounds right, yeah, yeah.

Fitzpatrick:

And that was a little before my time, so [laughs], unfortunately.

Richtmyer:

How old are you?

Fitzpatrick:

I'm 28.

Richtmyer:

Twenty-eight. A little older than my grandchildren. And you live in...?

Fitzpatrick:

Well, I live in Santa Fe, right now.

Richtmyer:

Oh, but you came from...?

Fitzpatrick:

I came from New York State.

Richtmyer:

New York State? Where?

Fitzpatrick:

Near, actually in between Utica and Syracuse.

Richtmyer:

Ah.

Fitzpatrick:

My family's farm is up in, actually —

Richtmyer:

Oh, on the farm, yeah.

Fitzpatrick:

Well, we don't, it's not a working farm, but I —

Richtmyer:

My folks both came from farms. They all came from farms.

Fitzpatrick:

We still have our original family farm in North Bay, New York, which is on the north shore.

Richtmyer:

But then you went to college?

Fitzpatrick:

I went to Virginia Tech, an engineering school.

Richtmyer:

Virginia Tech. Where?

Fitzpatrick:

It's the engineering school.

Richtmyer:

Where? In what city?

Fitzpatrick:

It's in Blacksburg.

Richtmyer:

Blacksburg.

Fitzpatrick:

It's way in the south, I mean, about four hours southwest of D.C., so, and then to Los Alamos.

Richtmyer:

And you've been in Los Alamos for several years?

Fitzpatrick:

Only a little less than three.

Richtmyer:

Yeah.

Fitzpatrick:

And I'm on the job market as well, so [laughs] there's very few academic jobs these days. It's very, very bad.

Richtmyer:

I know. Times have changed. In my time I never applied for a position in my whole life. They just came along, you know. Well, my father's influence was of course part of it.

Fitzpatrick:

What did your father do?

Richtmyer:

He was professor of Physics at Cornell.

Fitzpatrick:

I see.

Richtmyer:

And of course he helped me, but times have changed enormously. I never applied for a job, I never asked for a raise in my whole life.

Fitzpatrick:

So, well Hans Bethe is still a professor of Physics at Cornell.

Richtmyer:

That's right.

Fitzpatrick:

Still works, still teaches a class in it too.

Richtmyer:

He came just about the time I left.

Fitzpatrick:

Yeah, yeah.

Richtmyer:

But I got to know him later.

Fitzpatrick:

Sure, sure. Yeah. He's a fascinating guy.

Richtmyer:

He's still there, isn't he?

Fitzpatrick:

He's still teaching.

Richtmyer:

And he's officially retired?

Fitzpatrick:

Yeah, but he's always — I had a friend who was just in the Physics department there, just graduated and says he still comes in, you know, and I think he still has a few students, and you know. I think he teaches a class now and then. I don't know if he teaches regularly, but still publishing a few articles here and there, so some things in Physical Review.

Richtmyer:

And is his wife still alive?

Fitzpatrick:

I think so. Rose? Yeah.

Richtmyer:

Rose.

Fitzpatrick:

Yeah.

Richtmyer:

Yeah, yeah.

Fitzpatrick:

I think she is. So he's still a consultant to T Division, too. He was out two years ago, and I spoke with him then. Last summer, I think last summer he was going to Cornell, and then he decided he was going to Europe for the summer, and then, so he told his secretary, Sarah, it's presumptuous of him to assume he'll be here this summer, so [laughs]...

Richtmyer:

Is he in fairly good shape, physically?

Fitzpatrick:

Yeah. Last time I saw him, which was, you know, now two, three years ago, but apparently he's doing quite well. But to get back, to get back with one more — One other thing I'd like to talk about is Monte Carlo. There's been some material published on this recently, and some historians of science have been getting into this topic, including Peter Gallos and he's head of the History of Science department at Harvard. He's argued that Monte Carlo arose almost exclusively out of the super work. The fact that so many — He said that the fact that so many affects within the super had to be left out of the ENIAC calculation in 1946, and this is also about the time that Umlaum supposedly gets the idea, original germ of the idea, from playing solitaire.

Richtmyer:

No. Well, who knew? Stan Umlaum [?] was an imaginative person.

Fitzpatrick:

Oh yeah, and you were quite close with him, is that — pretty close friends?

Richtmyer:

Yeah, yeah, yeah. But he, at one point he had a delightful idea. A whole party of people, you know, playing games, but somehow all the things, records kept just such a way that get imitated, a multiplicative process.

Fitzpatrick:

Right. That's statistical. I mean, it was more of a —

Richtmyer:

Yeah. And his idea, he enjoyed this site [???], without ever proposing to make a very practical or anything, and then Von Neumann was out there at Los Alamos for a couple of weeks, and he caught the germ of the idea from Stan Umlaum, and then he went back to Princeton by train, he didn't like to fly.

Fitzpatrick:

Yeah, he was afraid of flying, yeah.

Richtmyer:

On the way back he wrote us a letter about the idea and how to program it through the ENIAC.

Fitzpatrick:

Which I've assumed that'd be this.

Richtmyer:

Yes. And this letter was addressed to me because I happened to be T Division leader at that point.

Fitzpatrick:

I see, uh huh.

Richtmyer:

And I got together a group of people, Metropolis, Umlaum, all those people, we talked it over. We thought it was a great idea, but we wanted to make a few little suggestions about it, so we wrote a letter back. I wrote it.

Fitzpatrick:

Which I think is also in here.

Richtmyer:

As far as I know, his letter to me and my letter to him were the first...

Fitzpatrick:

An iteration of that, or...?

Richtmyer:

The first document in the Los Alamos.

Fitzpatrick:

Well, this is a Los Alamos report.

Richtmyer:

Yeah, yeah.

Fitzpatrick:

Actually, its L.A. effect, you know, statistical methods in neutron diffusion, which has also been declassified. There were some histories done on Monte— Metropolis wrote a kind of history of Monte Carlo for Los Alamos science about nine or ten years ago or so, probably had this declassified in connection with that. One thing that's never— I've read this stuff, and one thing that's never really been clear to me is, now who was involved, who was working on this Monte Carlo? I mean, there's Umlaum obviously kind of originates the idea.

Richtmyer:

He originated the idea. But then when Von Neumann caught onto the idea, he wrote back and then various people got involved.

Fitzpatrick:

Yourself and...?

Richtmyer:

I was maybe [???] —

Fitzpatrick:

And Metropolis?

Richtmyer:

Metropolis and we used the ENIAC for some.

Fitzpatrick:

Well, right, because Von Neumann says that this sort of approach, the statistical approach is particularly well-suited to digital treatment, something like that.

Richtmyer:

Oh yeah, yeah.

Fitzpatrick:

And I was just wondering what kind of problems you might have had in mind when he was — I mean, is this to do, again, is it thermonuclear processes or is it to do with just regular fission, or...?

Richtmyer:

Yeah, well, the original idea had to do with neutron multiplication and diffusion in that system.

Fitzpatrick:

Right.

Richtmyer:

And I don't think it was more specific than that.

Fitzpatrick:

I see. And I know he wants to use the ENIAC for this, and this time there's really no other computer, I don't think, that the SSEC would be — this is —

Richtmyer:

I don't think they knew about the SSEC for a while. It's really —

Fitzpatrick:

Well, IBM kept all its machines secret anyways.

Richtmyer:

Yeah. So nobody had — but everybody knew about the ENIAC.

Fitzpatrick:

Sure. And, you know, according to T Division's reports, I mean, throughout '48 and '49, a lot of Monte Carlo problems, Monte Carlo technical problems, were taken back to the ENIAC.

Richtmyer:

To the ENIAC.

Fitzpatrick:

To the ENIAC.

Richtmyer:

Yeah.

Fitzpatrick:

Metropolis I think was doing —

Richtmyer:

Yeah.

Fitzpatrick:

I think, in fact, in order to do this, it's Metropolis and Corey Von Neumann, that they actually tried to convert the machine to a stored-program computer, because initially it wasn't a stored-program computer.

Richtmyer:

That's right. That's right.

Fitzpatrick:

And then, I think, they were just —

Richtmyer:

But they built in a program, a memory at some point. They actually had 200 words or something.

Fitzpatrick:

Yeah. [laughs] By our standards that it's not much. But at the same time — but they, I know there was a lot of problems sanctioned to go to be run on the ENIAC from T Division, you know. It was like sometime after this is all proposed. And these have to do with just a variety of problems, and including, I gather, implosion. There were some problems on hydroid weapon, you know, it was this hydroid weapon that Hyman and others worked on during the war. It was kind of dropped, and then, you know, picked up again for a while, and then it was apparently so inefficient that it never worked. There was a bunch of — But Monte Carlo is applied to all kinds of different types of problems.

Richtmyer:

Oh yeah, yeah, yeah. Yeah, it was a very general sort of method.

Fitzpatrick:

Um hm. Because I think later, I think much later, in the summer of 1950, Umlaum and Cornelius Everett, both at Los Alamos, or Everett's there at least for the summer, and by this time there's just a lot of discussion, more discussion about Teller's super after Truman's given his directive, I think, to continue with thermonuclear work, and they start to do hand calculation, I think very simplified Monte Carlo techniques on the ignition of the super. And this time, they come up with a result that probably won't ignite with less than a kilogram of tritium. And this is, you know, I think what starts to kill the thing. Later, I think that someone runs this thing on the ENIAC, which, you know, there in '53 or '54; it's much later, and confirms the thing. So this is applied to all kinds of problems. So it's also — but it is. It's just a series of people. That's what's not clear in the historical literature, you know, more due to Von Neumann, Monte Carlo, to just Von Neumann, or Von Neumann and Metropolis, or Von Neumann and Richtmeyer, or, you know.

Richtmyer:

Yeah, well the original idea was clearly Umlaum. No question about that.

Fitzpatrick:

He's written about that in his memoirs, too.

Richtmyer:

But when Von Neumann took hold of it, then that's when it began to really be significant.

Fitzpatrick:

Well listen, skipping around again. When you started using the UNIVAC machine in 1953, was it that late, or was it before that?

Richtmyer:

I think it was before that. We used the one in Philadelphia.

Fitzpatrick:

Right, because you were still using CAC as late as '52, I think, so, and then you moved from the CAC to Philadelphia.

Richtmyer:

To Philadelphia, yeah. Because that was a somewhat bigger machine.

Fitzpatrick:

And probably a faster machine, too.

Richtmyer:

Right, faster. Yeah, yeah.

Fitzpatrick:

And, but are you still running the same alarm clock calculations on that, do you remember?

Richtmyer:

I don't remember it at all, what we ran. I'm sorry to be so —

Fitzpatrick:

No, this is very helpful.

Richtmyer:

I remember using that machine in Philadelphia and I remember —

Fitzpatrick:

Who did you use it with, do you know?

Richtmyer:

Well, some of the employees of the company would work with us, actually using the machine by hand and all that.

Fitzpatrick:

Right.

Richtmyer:

Although, we learned how to do those things, too. But we had a nice relationship with those people. We used to go out to dinner with them and have drinks.

Fitzpatrick:

Well, a lot of those early programmers are fascinating. I wish I could track some of them down, but it's usually hard, because —

Richtmyer:

I don't even remember names of them, any more. But on the Hippo thing, I think it was Clara Von Neumann and Adell Goldstein. They spent hours on it.

Fitzpatrick:

The family got to be very good programmers.

Richtmyer:

Oh yeah.

Fitzpatrick:

All kinds of problems, so.

Richtmyer:

And when we started using the SSEC, some of the IBM people took me and said that some of the things that Adell was doing were much too complicated, but they worked.

Fitzpatrick:

Yeah. Well, I read that, I mean, IBM's people had a real problem with Hippo, and had a terrible time with the machine in preparation for Hippo, I think, because they'd never seen a program like this. It took them a really long —

Richtmyer:

Yeah, well, we did all the work.

Fitzpatrick:

I see, uh huh.

Richtmyer:

The machine, a great big machine with lots of engineers to help us and all that.

Fitzpatrick:

Yeah, but I think that Hippo was delayed partly because IBM's people weren't familiar with —

Richtmyer:

Well, they weren't the ones who did it. It was largely Adell Goldstein who programmed the whole thing.

Fitzpatrick:

Were you fairly close with the Goldsteins?

Richtmyer:

Oh yeah. And in the early days of the Hippo, Adell Goldstein and I shared an office in that IBM building.

Fitzpatrick:

And Clara Von Neumann.

Richtmyer:

And Clara Von Neumann to some extent, but Adell mostly, and we had those rooms, you know, 24 hours a day, but at that time I was playing viola, and I used to practice my viola sometimes at night, and the watchman thought this was all very strange.

Fitzpatrick:

Yeah, I imagine so. These were the offices on Madison Avenue; this was IBM's headquarters, right?

Richtmyer:

Yeah, that's right. 590 Madison Avenue. I think I've got the number right.

Fitzpatrick:

Sounds right to me, yes. Sounds exactly right. So then you moved from SSEC to CAC, was the next machine you used?

Richtmyer:

Well, CAC was mostly for somewhat smaller problems.

Fitzpatrick:

And as I recall, it was just alarm clock. It was just the alarm clock. I don't know if it was used —

Richtmyer:

Well, the SSEC might have been used for some of the problems we had, but it really took the, I'm sorry, the CAC. The SSEC was the big machine, enormous, slow and clumsy, but it was big.

Fitzpatrick:

Which is why it was used, instead of ENIAC, as I understand. I think that's about it. That should do. All right, thank you very much. And for the record, which I never stated in the beginning, we are here today at the home of Robert Richtmeyer in Boulder, Colorado. Today is March 4th, 1997.