This section of the Desch/Mumma interview runs from page 124 to 166 of the first day transcript
[Page 123 cont.]
RM: .. Y’ see I was more application oriented and he was more theoretically oriented and that was the big difference
JD: Mhm. That’s right.
RM: between the two of us.
HT: Mhm. You say he eventually became the scientific advisor to the President; and when was that?
JD: Nixon, you bet. Now, why he left, I don’t know. But I noticed that when Nixon entered the White House for the first time, that is, the first four year term, .. Louie DeRosa’s name appeared in the press as a technical advisor to the White House, and, now I can’t prove that, it’s just-
[Page 124]
something that I, seem to tuck away, because he didn’t work for me anymore. But-
HT: Mhm.
JD: and I don’t – I don’t see him anymore, but it was one of those things that made me very happy. I was quite proud of him
HT: Mhm.
JD: to think that he could make the grade and get up there.
HT: Did you say he was no longer alive?
RM: Yes, he’s dead. He’s been dead a couple of years –
JD: That shocks me because I didn’t know that.
RM: Yes, I saw it in some type of a magazine, Joe.
JD: Is that so?
RM: He died more recently than Frank Bucher, and Frank Bucher died several years ago.
JD: Yeah, Frank Bucher worked for me, too.
RM: Frank was at Bell Laboratory.
JD: Yes, he went to Bell Labs.
RM: I thought that DeRosa carne from Bell Labs here.
JD: I don’t recall. I can’t recall where he came from I’d have to have those resumes –
RM: Yes, I do not remember, he was here when I came, I know that, because DeRosa and Killheffer were both here.
[Page 125]
JD: I had four or five men here when I got Bob. Bob came from Frigidaire, and – and picked up this concept that I had written up earlier, and with modifications and additions and so forth, he essentially followed that pattern. And, of course, this was the subject of a violent discussion in the court room
JD and HT: [Chuckle]
JD: when the lawyers tried to prove that – that he was the total inventor and I had nothing to do with it, see.
RM: But it was a joint patent.
JD: And so I don’t know how it ended. Did it end that there was joint interference?
RM: Yes, the first model was a joint one. On the one that had the calculator, that I was sole inventor on that one, but the first one that we went into interference with IBM, why we were joint – you were joint on that.
JD: Yeah, but there was another one. They – they switched that thing three times – or twice.
RM: Well, they split the case – I mean, they broke it up into sections, you see.
JD: Yeah. Well, anyway, I – I noticed that the – that the court .. made one statement in its decision
[Page 126]
and in their opinion they had serious doubts that I was even an inventor. [Laugh]
HT: [Laugh]
JD: And they were using my notebook all the time.
HT: [Laugh]
RM: Well, of course, I was more closely associated with it and I knew the details.
JD: Well, we communicated daily.
HT: Mhm.
JD: I mean I had not over 20 people, and I got to talk and sit down with all of them every day,
HT: Mhm.
JD: with no communication problem at all. I knew exactly what everybody was doing every day that I was here. And so the lawyers tried to make something out about this – my lack of communication and that was a lot of baloney because if it wasn’t in that book, I was telling them what I thought. And while I didn’t order anybody to do anythingI never did because it was a semi-research thing . and I knew you don’t order people to do research. .. Nevertheless, you guide them, even express an opinion and, in the end, they’ve got to have something that works.
[Page 127]
HT: Right.
JD: So the ultimate test
HT : Yes.
JD: is not whether they followed your orders but whether or not they made it work.
HT: Ahm. Well, back on the pumping station, .. .. maybe we ought to trace that through. Because you showed the diagrams, approximately 1948, ’49, as it existed if we were to walk over to Building 20. today, but you were building these small thyratron tubes .. almost immediately when you built that first –
RM: 19– 1940.
HT: Right, that first prototype,
JD: Sure.
HT: which means that you had to have some kind of an operation like that pumping station.
JD: Yes. There is a diagram there of the very experimental station.
HT: Mhm.
RM: Have you got your number — Was that in your number one, early notebooks, or not?
JD: Ah – ..
RM: Well, I don’t know.
[Page 128]
JD: I – it’s all –
RM: Of course, it might be — see we have Cone’s notbooks around here too which might show some of that.
JD: Yeah, but you find them.
RM: Well — why, I
JD: Well, I talked to them again today to try to find out those other notebooks.
RM: Well, I don’t see my notebooks here and there are a big stack of mine around someplace.
HT: Right. I haven’t seen them.
RM: I saw them the first of this year because I went through them and picked them all out. .. .. ..
HT: I found the – the deposition.
RM: – I could get the notebooks for you if we could find them.
HT: I’ll turn this off for a second.
[Recorder off]
RM: This third calculator that I described earlier, had the layout showing how I was going to divide on that calculator, but I never got it checked out. All of the thyratrons for division were installed in this calculator.
HT: Now, this is the one that you demonstrated during your recent litigation?
[Page 129]
RM: Yes, that’s right.
HT: And, everybody was vague when I asked them about division and apparently that’s the reason why. That it really hasn’t –
RM: It was never made to work. All the tubes were there and the principle of its operation is described in one of these notebooks. In fact, I even wrote the description of division in this calculator for the purpose of the trial.
HT: Aha.
RM: How it really operated. I picked it out of the notebook.
HT: And this is “division is planned for the NCR computer, patent model number 3754,” and this .. three page description is how it would have operated had you decided to
RM: Yes.
HT: complete it in that particular model.
RM: That’s right.
JD: You see, the war interfered with that or you would have had division.
RM: That’s right. Honeywell was being sued for patent infringement by Univac, and they wanted me to make this model work.
[Page 130]
HT: Mhm.
RM: I had a horrible time because of starting with this little blueprint and the 30 year time lapse.
HT: Yes.
RM: I found more thyratrons than were shown on that blueprint. I couldn’t understand how the thing worked until I finally got back to my notebooks or back to some drawings, I finally got the original drawings for this electronic calculator, which I used.
HT: Mhm.
RM: And found out that I was going through some tubes that had no function. They were switching tubes. They were there for switching from multiplication to division. They were not necessary for multiplication.
JD: The division wasn’t there.
RM: Yes. But, the tubes were there and I had to go through the tubes and use ’em and switch them the’ proper way.
HT: [Laugh]
RM: But we finally got the machine running. They did demonstrate it at the trial in Minneapolis. It ran, it added, subtracted, and multiplied.
[Page 131]
JD: You did a wonderful job in demonstrating that in the courtroom for the judge.
RM &
JD: [Laugh]
JD: Talk about demonstration.
HT: I’d like to go back and pick up the chronology of these .. thyratron tubes, because the notebooks that you just showed me with the full pumping station, with your final modifications, they’re approximately ’49, when you finally get the purification and all the gauges and everything running the way you want it, .. but you were manufacturing thyratron tubes almost a decade earlier. Now that means that you were making them in this building. [pause]
RM: Yes, third floor.
JD: A large quantity of them, yes.
HT: In this building, under – using very different equipment.
JD: Well, simpler equipment. And it was one of the reasons why we didn’t get uniformity until I improved the pumping station sufficiently that we controlled the variables;
HT: Mhm.
JD: that is, part of it. .. We could — we, we knew
[Page 132]
where we were on vacuum; we knew where we were on purity of gas; we knew where we were on pressures; and we also learned a lot about activating cathodes And also how to prevent the leakage across the micas in the tubes. And when we got all those things straightened out, then we built tubes in quantity.
HT: But still in this building?
RM: Yes, that’s right.
JD: But now, there are many versions of the pumping stations that you have no photographs of here at all
HT: Right.
JD: that I kept changing it and changing it and we never photographed it at all as we went along,
HT: Mhm.
JD: until I got it across the street into Building 20 and I was relieved of my duties in the
RM: Administrative.
JD: electronic research and Bob Mumma took charge of electronic research and I worked alone over there. And then I was able to put in the time that I wanted on inventing new tubes and to build a real
[Page 133]
decent pumping station which, at its – at its time, was one of the best in the country.
HT: What was the – the main use of the tubes you were turning out in 1949?
JD: Those devices that I showed you in the – in that book over there, were all that were – wherever you saw a multiplicity of those small tubes; those – those were the tubes we were making. And the .. .. and the .. .. and the storage tubes and some ten digit counting tubes.
HT: Were you also supplying tubes outside of the company?
JD: No.
RM: No. Wasn’t your goal then, Joe, to develop a cold cathode tube that could be made very – very cheaply?
JD: Yes.
RM: And we could make a computer for business appliction where the speed didn’t have to be so high
HT: Mhm.
RM: and make something that could use these cold cathode tubes.
JD: That’s what I was doing over there across the street when I was alone.
[Page 134]
RM: That’s right.
JD: And.. .. and that’s what some of those books were devoted entirely to — all kinds of concepts.
RM: Because you – you tried all kinds of materials .. zirconium or some of those materials for cold cathodes to get .. .. lower drops, you see –
JD: Yeah. All that work is documented and countersigned by either Ed Herman or Louie Kline.
RM: Yeah, I expect it is.
JD: All of it is. And – and all of the concepts were tried. I made the tubes.
RM: Yeah.
JD: I had everything right there to make it with,
RM: I know you did.
JD: and I made the tubes. And the counters worked, because we got patents on them. There’s a half a dozen patents
HT: Mhm.
JD: in my name on those cold cathode counters that he’s mentioning, .. .. but – but their speed was not sufficiently high. Again, it was speed that had me blocked because a cold cathode tube deionizes very slowly. In fact, it fires very slowly; it ignites very slowly; so to make it go on and off,
[Page 135]
and on and off fast enough, you couldn’t — well, to get 3 or 400 impulses per second you were doing real good.
RM: I don’t think you ever – you hardly got to a thousand, I don’t believe, did you?
JD: No, I didn’t, hardly.
HT: Well, what you were primarily concerned with then was essentially the same problem that you were concerned with in 1938?
JD: Right.
HT: To convert your product line to electronic as opposed to mechanical maintenance.
JD: Right.
HT: And you weren’t thinking then in terms of any kind of a programmed operation. You were thinking primarily of the accounting posting machine, the cash registers.
JD: I was doing the same thing that MIT was doing. I was still hunting for the ideal counter of – thof minimum cost
HT: Mhm.
JD: with a respectable speed that would enable it to be useful. That’s what I was still searching for.
RM: Before the transistor came out.
[Page 136]
JD: Before the transistor –
HT: Oh, yes, that’s still a few years before the transistor.
JD: Yeah, but it wiped me out. I still had that running when the transistor hit, and. then I closed down.
HT: Well, when the first transistors hit, of course, nobody knew what they were going to do. Their characteristics were totally unknown. They were unreliable. They were very expensive, and it was really — nobody expected them to be as good as they were.
JD : Well, ..
HT: They were a good example of something that far outlived and outraced the expectation.
JD: Yes, I – I had my doubts about certain things. But when it became apparent that the transistor was going to go — I tell you, it occurs, and in a – and in a little while, this afternoon, you’re going to find out when it occurs is when we go into the 304 class computer that we built .. which – which was developed at Hawthorne, California.
HT: And this is after the acquisition of CRC.
[Page 137]
JD: After that acquisition.
HT: Yeah, before we get into that, you –
JD: But – but, b- but in between there, .. .. we’d, we – of course, the transistor became very important. But I want to make this statement on the record, before I forget it. NCR offered the first commercial, all transistorized computer on the market. If you don’t know it, you know it now.
HT: Mhm.
JD: And very few people know that.
HT: Okay, would you care to id-
JD: We offered for sale and delivery the first transistorized computer.
HT: Would you care at this point to identify name, dates?
JD: I can’t – I can’t do that. I – see, I didn’t know it myself at the time.
HT: Aha. Who would be the best person to – identify?
JD: Eckdahl.
HT: Eckdahl.
JD: Eckdahl would be the best person.
HT: Okay. Well, Don and I during our conversations totally concentrated on an earlier period, wh~ch was the – his involvement with Northrop and the origins of the CRC.
[Page 138]
JD: Yeah, well, we will talk about the 304 a little later because Eckdahl is involved in it and so are we.
HT: Mhm.
RM: That was the name of the first transistorized computer, the 304.
JD: Internal program.
HT: Mhm.
JD: And, by the way, we still have a flock of them out. They’re still operating.
RM: That was built for us by General Electric.
JD: Well, part of it was — the computers
RM: The main frame.
JD: and some of the –
RM:- The main frame.
JD: some of the buffers.
HT: Well, you mentioned, before I turned the machine on when we came in this afternoon, that you were a long time corning around to the concept of a stored program computer. And you might want to elaborate and – and talk about that a little.
JD: Well, after we were visited by [clears throat] four or five men from CRC — and I can’t tell you the dates of that but there ought to be something
[Page 139]
in the records somewhere — Eckdahl was one of the members of the team and – and – and president of the company, CRC .. was also a member, and, and a guy named Sprague was the sales manager,
RM: And the financial.
JD: and another man was Warschau was the
RM: Financial man.
JD: treasurer, the financial man, and I forget the president’s name.
RM: I can’t think of the name either.
HT: Yeah, I was going to say Floyd Steele had been, but he had been replaced by then, by a gentleman who had an accounting business background.
JD: Yeah, well, I know his name
HT: I’ve forgotten his name.
JD: as well as mine, but I can’t think of it at the moment.
HT: Yeah .
JD: I – I may –
HT: Harold Sarkissian, was he part of that group?
JD: Sarkissian was the engineer in charge at the time.
RM: He was the overall engineer, yeah.
JD: Well, he wasn’t in that original group that came.
RM: I know that he didn’t come to NCR but his …
[Page 140]
JD: So, they carne in and we held a meeting in the boardroom downstairs and they presented their case and they were in financial trouble. There wasn’t any question about it. And I had been fighting for several years to make an acquisition because it was becoming apparent to me that we had to get into this area of business, at least for two years before this acquisition. One was to purchase UNIVAC, which I told you previously, I – I made a strong case for that, and actually we started negotiations when somebody snatched it away from us
HT: Mhm.
JD: and that was Sperry Rand.
And then I got a telephone call one day from ERA of St. Paul because it was populated with all of my old friends from the war.
HT: From OSRD, that’s right.
JD: And they – they told. me, they gave me the whole balance sheet, and all the business, and all the orders, and the financial statements, everything about the place, for $4.50 a share. So I wrote a letter immediately and in 15 minutes I headed up to the president’s office and he turned me down.
[Page 141]
So I had to call back and tell them no and they said, “Well, it’s just as well. Sperry Rand just bought us.”
HT & JD: [Laughter]
JD: For $5.00 a share.
HT: Well, the question
JD: Did you know – did you know Engstrom – fellow by the name of Engstrom?
HT: No. I know of him.
JD: From ERA?
HT: The only cont-
JD: I don’t mean the RCA Engstrom.
HT: No, I know. And the only contact I’ve had directly with ERA — I’m trying to think of his name now. He’s one of the co-authors of their original publication on computers.
JD: Oh.
HT: And he – when I talked to him –
JD: Do you – do you know Bill Norris?
HT: Yeah. But this gentleman was Assistant Secretary of Commerce.
JD: Well, he was at ERA, you know, in a vice presidential capacity.
[Page 142]
HT: Right. Excuse me.
[Recorder off]
JD: It’s apparent that by 1950 or thereabouts I was convinced that our approach to electronics .. .. would .. ..really not compete eventually with internally programmed computers, and that is the reason for my attitude that we would purchase a company that did have a suitable computer and – and break into the – that type of operation by purchase of a corporation that needed financial support. And the first one was – it was the UNIVAC I, Eckert-Mauchly Corporation, and the next one was ERA, the Engineering Research Associates. And neither materialized, and then .. .. some time later, not too much later, but perhaps in ’51 or thereabouts, the record will show exactly when, we were approached by people from CRC who came to see us. The meeting was held in the boardroom and their presentation was made. .. Of the people in that room, .. only two of us supported that purchase because Mr. Keenoy, my superior at that time and the head of the product planning – product development – department, and myself were both convinced that we had to do
[Page 143]
something about getting into the .. .. .. business of a .. internally programmed computer. And – and, and we were in the process at the time of writing the specifications of a computer, internally programmed, which we were going to build. Mr. Mumma and I were working with product planning on this question.
And I made the case .. suttic1entLY strong that the management of the corporation decided to take a look at CRC. So we went out to California and we looked them over. Mr. Mumma and I, and Mr. Keenoy and the Los Angeles Sales Manager.
RM: S. C. Allyn was out there.
JD: .. Yes. S. C. Allyn arrived after we had spent – a day or two there ahead of him.
RM: That’s right, yes.
JD: And the manager of our sales division arrived. I think Mr. Oelman came out, too. I’m not quite certain anymore.
JD: After Mumma and I had looked them over technically we made the case that we should buy them. And, of course, we didn’t enter into any of the financial arrangements, but we did recommend that
[Page 144]
this was a – one of the last cases they were going to have of acquiring a corporation that had already developed a computer.
HT: Mhm.
JD: The computer they had was scientifically oriented,
HT: Mhm.
JD: and it had to be modified. And we had several applications to make it suitable for commercial applications.
.. After some discussion back here at Dayton, and some meetings, it was finally decided to – to purchase them and we did that. .. The chairman of the board backed us up and we purchased them. And then set about to try to convert their equipment by – through specification changes and so forth, in to. something that we could sell. And that’s how we got into the internally programmed computer business.
HT: But you were thinking then in terms of buying a ready going concern?
JD: Yes.
HT: Of buying the talent and the expertise so you. wouldn’t have to duplicate all of this that had been done —
[Page 145]
JD: That’s right. There was no sense in spending all that money to duplicate something that we could buy.
HT: And you weren’t as much concerned with selling .. scientific computers as with the business applications
JD: That’s right.
HT: and hence the modification of these for business users?
JD: Although ironically we do sell for scientific applications now,
HT: Mhm.
JD: but that came a little later. Originally we didn’t.
HT: You mean like the CRC, .. what did they call it? The 101, that was the digital differential analyzer. Now this wouldn’t have been then a product that you would have been interested in manufacturing?
JD: No, we were not interested in the differential analyzer.
HT: How about the 102, which was their binary
JD: That’s the one
HT: computer.
JD: we were interested in. The 102 and later the 103. And those were the ones that we were really interested in to convert to business applications.
[Page 146]
And those, primarily, were to be used in the analyzation of tape produced by cash registers and accounting machines.
HT: So you were thinking then in terms of what is finally a realization today?
JD: Today we now have terminals that go to computers
HT: Right. But-
JD: and that was in our mind back in 1951, ’52.
HT: that was a dream then and apparently I guess the only conclusion today is the technology wasn’t quite up to the dream.
JD: No, the speed of the computer wasn’t great enough at the time. Speeds are now pretty high; and I notice another thing: There’s another aspect of our thinking and that was that we were somewhat oriented to specific application machines, like we were with our accounting machines. They were built to do certain things for certain people and not to try to do everything for everybody.
HT: Mhm.
JD: Well,.. .. in the beginning that was a very difficult thing to weed out in that it was now possible to build a self-programmed, an internally programmed, computer that would, would by simply
[Page 147]
putting a different program on the machine, do an entirely different job.
HT: Mhm.
JD: And that was a new, a brand new concept to NCR, and had finally been accomplished.
HT: Well, I noticed .. you can correct me because I’m vague on details: I was going through a chapter, I think it’s in Allyn’s book on his years with this organization. .. There’s a chapter on computers, and he mentions the suggestion to .. get the banks interested in this kind of equipment for their purposes, and the great antipathy on the part of – of management that they would ever convince a group of conservative bankers to look at a whole new system and a whole new way of doing things. You might want to elaborate and expand on that.
JD: Well, I can elaborate a little bit on that.
HT or RM: [Cough]
JD: There existed within the corporation — it was – it was one of those things that’s hard to explain who did what, but the question of putting accounts on magnetic tape that could be erased, or tapes could be mixed up and you could get the
[Page 148]
wrong account number and get somebody else’s account, and so forth.
HT: It still happens. [Laugh]
JD: Yes. But that was so strong.
HT: Yea.
JD: And, also it was a question of the legality of maintaining records on tapes, financial records, that there was a question as to whether or not it was even legal at that time. Now, Mr. Kettering, Charles Kettering, .. was proposing the use of wire records at that time, but we never followed the wire recorders. But, he wanted to do it, too, and he – and he practically owned Winters National Bank at Dayton, Ohio, and – and he was willing to – do it. But,.. .. .. the, the question now no longer exists, this question of mixing accounts, and getting the wrong account numbers and so forth; and if you really go back to it, even on the mechanical machines, it’s easy for a girl to push the wrong button.
HT: Of course. But, what kinds of inertia had to be overcome within NCR during that period, because, as you say, the whole thinking had to be changed?
JD: Well, it was a slow process. .. There had to be
[Page 149]
test applications. We had to – we had to put in machines in a lot of places and test them out, proceed with caution, until we were sure that it was safe to do it. Meanwhile, [there were] other people that were doing it, but there were other people getting into mighty serious trouble, too, in some places where they jumped the gun a little bit too fast.
HT: Mhm.
JD: But – there had to be safeguards. You see, we’re a company where when we build a cash register or an accounting machine, we put in interlocks and all kinds of devices to prevent making mistakes.
HT: Mhm.
JD: And, in cash registers, to prevent the robbing of the cash register.
HT: Mhm.
JD: That is, you just can’t open it unless you do cetain things, and the same thing with accounting machines. If you – if you push the wrong sequence of events, it won’t work. Well, the – the question of doing something like that on .. on these, on these self-programmed computers, no one knew how to do it. We didn’t know how you were going
[Page 150]
to check. So, so systems involving .. .. account numbers that were – that were determined mathmatically, that certain account numbers could, would not function in the machine. I forget what they call that, Bob — we built machines that do that.
RM: It’s a self-checking account number.
JD: Yes. You got an account number and if it wasn’t entered correctly, the machine wouldn’t work, and that means that there – there was a limit to the number of account numbers that we could use because we had to throw out a lot of them.
HT: Mhm.
JD: .. We built machine to do that.
RM: See, the last digit has some bearing to either the sum of the other digits divided by nine or some, .. system was worked out.
HT: Mhm. So you had an internally checked digital approach.
RM: Yeah. And some were better than others for getting interchanged digits and – it’s in common use today f on all these account numbers in banks, this check, ing system.
HT: Mhm. So that a simple permutation of two or three
[Page 151]
digits then would cause things to just stop.
RM: Yeah.
JD: Well, they were longer than two or three digits.
HT: Well, I mean if – if I was putting in an account manually and I permuted –
JD: Oh.
RM: And permuted or interchanged, that’s right.
HT: or interchanged two or three digits,
RM: That’d be an “Error.”
HT: things would just stop and I’d have to go back and say “uh-oh, I copied this wrong.” .
JD: Yes.
HT: Now, let’s go back to my original question which .. was going to lead up, I – I think, to something else. How did you yourself come around to the idea to abandon your original approach?
JD: How did I do it?
HT: Yea. You mentioned when we said, you know, when you were working on your.. ideas in ’48 and’ 49 that you weren’t convinced —
JD: When we purchased CRC, the die was cast as far as I was concerned.
HT: Yeah, but you obviously had to be convinced earlier or you couldn’t have sold other people.
[Page 152]
JD: Well, I said, during the two years preceding the purchase of CRC, I was already convinced.
HT: Mhm. Right.
JD: Otherwise I wouldn’t have recommended,
HT: Right.
JD: but with some urgency,
HT: Right.
JD: the purchase of UNIVAC I, and also to get in to buy ERA.
HT: Of course –
JD: That took nerve on my part to even recommend that, because I was standing alone on these — well, on those two I was standing alone.
RM: Of course, IBM at that time was moving in that direction, and –
JD: Not till about 1951, really.
RM: Mhm, I see. Yes.
JD: But on, on – on the case of the CRC I had a partner and that man today is a vice — group vice president of sales.
RM: Sales.
JD: Sales – Chuck Keenoy and I stood together and WE held, because we both realized that that’s what we had to do. That’s all there was to it.
[Page 153]
HT: Well, let’s go back .. .. .. into that .. convincing period of yours, because you mentioned that you and Bob laid out the specifications for an internally stored machine.
JD: No, we did not.
HT: Oh, I’m sorry. I misunderstood then.
JD: We – we received the specifications from Chuck Keenoy and his .. .. righthand man of the specifications of a machine, question mark, of a suitable design
HT: Mhm.
JD: to accomplish certain end accounting results, and it was to be electronic, perferable. And – I suppose it was dictated to be electronic for speed. .. But, it – it never was – it never was, dictated that it would be an internally programmed –
HT: Aha.
JD: It – we could throw a switch and move it around to all kinds of things, and we can do that on any programmed computers. I
HT: Mhm.
JD: Just switch in different combinations by hand.
HT: Mhm.
RM: Joe, wasn’t it getting around to the fact that every time we turned around or something they would come up with another requirement for our
[Page 154]
NEAM, National Electronic Accounting Machine.
JD: That’s right.
RM: And I think this is when they became – [it] became apparent that we had to have something that wasn’t –
JD: Well, they were looking for a universal machine
RM: Yeah, that’s what –
HT: Mhm.
JD: and we couldn’t give ’em a universal machine with our approach. Bob is right. That’s exactly one of the things that drove us into the .. internally programmed computer where we could change the function of the machine easily. And, and whwhen we, when we realized that that’s what the company wanted because Mr. Keenoy was at a high level in the company even as the director of
RM: Future demand?
JD: future – future development organization, we listened carefully because we wanted to give the company what they wanted. And when we got ahold of CRC we had a large part of what we wanted, and that’s why I wanted to get CRC, was to get that ability And, then, of course, we had to modify the machine which took some time. And
[Page 155]
it ended up in developing a whole new machine as a second generation of machines for commercial work, and that was the one that was known as the 304.
HT: This is the transistorized?
JD: This is a completely transistorized machine. What Bob and I were working on, up until that time, until I completely stopped the work, was a vacuum tube approach. And it became apparent that we were going to be obsolete before we ever hit the market with tubes, so we – we switched to transistors, which set us back about six months. But we made it and we still got on the market before anybody else got on the market.
Now I don’t say that no one else had a prototype or –
HT: Mhm. Yeah, you’re talking about a production model machine
JD: Production model.
HT: that somebody could walk in and buy.
JD: Yes. We could sell it.
HT: I guess one of the – the curious things – in looking back I realize why you did what you did in terms of buying talent and experience in the
[Page 156]
field, but – given your past record of having entered fields and started from scratch, .. had you and Bob and other people in the company thought about a development program that would have led to an internally stored computer as a prototype development of its own?
JD: I can’t say that we did for the simple reason that we had just come out of about six, seven years of military development, entirely unrelated to computers, excepting what you saw on OSRD and NDRC.
HT: Mhm.
JD: But the Navy program was entirely divorced from computers.
HT: Mhm.
JD: – We here, in effect, were only about a year into reopening our programs again
HT: Mhm.
JD: and – and, and we hadn’t been thinking along the lines of computers during that period. And, as soon as we got into the business again and – and saw what the status was, then we became convinced that we –, and I had to make a plea to the chaman of the board — that we were on the wrong track and we had — and I – I convinced him that
[Page 157]
we were on the wrong track and that we had to reverse our field. So, that was – that’s about as good as I can say it.
HT: Oh, I think that’s very good. ..
I’m going to really ask a wild one now, andand that is, as you look back to your early tube works, beginning in ’28 and these early prototypes that you built, what do you see as their impact on the technology, state of the art, or developments in general? Do you see any role that they played?
[Pause]
JD: Well, as, as, as devices they played a role in that they acquainted us with small systems and things of that nature. But, as far as the technology is concerned, that has been completely wiped out, as far as I’m concerned, by micrelectronics and the .. large scale integrated circuit chips.
HT: Mhm.
JD: You’ve got the whole computer on a one inch square piece of silicon. .. I don’t see any connection at all
RM: No.
JD: except the general thing that we – we had a little
[Page 158]
preparation ahead of time but not an awful lot.
HT: You were going to tell a story later about the, .. when you realized the – what the transistor was going to do, and this might be a good place to do that, in terms of your shift that you mentioned just a few minutes ago, how that occurred to you, and what –
JD: It occurred because — you see, it took ten years after the first announcement of the transistor until it became a useful device in the computer.
HT: Mhm.
JD: It – that really is true,
HT: Right, right, right.
JD: about 10 years, because I had the original article. It was a letter to the editor in the Physical Review. It was the first blurp that came out on the transistor.
HT: Do you remember the approximate date?
JD: N- no.
HT: Within a year or so..– ’48, ’49, ’47?
JD: Oh, golly. I don’t know.
HT: I know it’s roughly in that time period because by ’50, ’51 people had a few expensive ones that they used to keep in the safe. [Laugh].
[Page 159]
JD: Had a wire whisker on it like the old silicon radio detectors.
RM: Yes.
HT: Remember they used to cost about $15 each and they were highly unreliable and full of imperfections.
JD: Well, you’d just jar the table and they’d go out. But, that – .. .. .. it took 10 years for it to become a practical thing.
Now, just where along the line I decided that it was going to be practical, I can’t tell you, but I was watching the developments very carefully. And I .. .. tried to evaluate it as best I could, as to whether or not it really was as good as I thought it was, and so we started to experiment with it and found out that it was a reliable device.
HT: Of course –
JD: Now, by that time it was one of these, not the cat whisker type, but it was the
RM: Silicon junction type.
JD: junction type transistor. And that put the – put the reliability into it.
RM: Of course in computers they used transistors opating at saturation current. I mean, they were either completely conducting or completely off.
[Page 160]
They did not use linear devices, such as those used in linear amplifiers. Other than switching speed, I think the requirements were less for computer applications,
HT: Yes.
RM: for transistors in a computer, where they are used in
JD: Well, you see
–
RM: switching applications.
JD: you can almost make a comparison between that and a thyratron,
HT: Mhm.
RM: Yes.
JD: especially the SCR’s.
RM: Well, they’re the same thing.
JD: That’s the same thing as a thyratron.
RM: Yeah. It didn’t require anything else to hold them on.
JD: You’d turn it on and it would stay on.
HT: Mhm. Y’ know, this, this square pulse was, of course, everybody’s goal. At Northrop they had a little project going with neon tubes because it looked like they ought to behave that way except that they didn’t. [Laugh] The curves just gave
[Page 161]
that kind of – of feeling. It never really worked out.
RM: That’s why Joe worked on cold cathodes because neon tubes were cold cathode tubes.
HT: It sure looked like that ought to be the way to do
RM: Yeah.
HT: when you looked at their characteristic curves.
JD: There’s an awful lot of work that had been done on – on cold cathode devices for a variety of things. But it never, we never realized until just a year or two before we practically abandoned all desire for all the use of neon gas, and argon gas, krypton and xenon and all the rest of th. we tried ‘ern all, we had all kinds of rare gaswas the fact that it’s a statistical device. Whether it fires or not when you put the starting voltage on it depends largely whether or not a cosmic ray hits it or some – something triggers it off
HT: Mhm.
JD: and it doesn’t always trigger off at the same voltage.
HT: It’s kind of an uncertainty principle. [Laugh]
[Page 162]
JD: It – it’s a fact. It’s a statistically operated device and – and that can get you into an awful lot of trouble.
HT: Right.
RM: We used to put fluorescent lights in – in cold cathode counters … so they’d always start.
HT: [Laugh]
RM: We had to have – we had to have some emission,
JD: We had to put fluorescent lights in there to illuminate ’em,
HT: Mhm.
JD: be sure we got some electron emission before, so we didn’t rely on a cosmic ray to trigger them.
RM: The ultra-violet light from a fluorescent lamp is enough to trigger emission at the required control element voltage.
HT: Yes, I remember somebody telling me about one of the early machines that was being built. And they discovered that things went very haywire on the nightshift when somebody left the machine running and turned off the neon overhead lights.
RM: Yes, that can.
HT: [Laugh]
JD: We found that, too. [Laugh]
[Page 163]
HT: You know, that they were very dependent upon just that very small emission from the overhead neon tubes.
JD: Yeah. Unless you greatly overload it in voltage, in which case you’re going to get very high currents, then you can insure some electron emission, but it takes a – it’s not just the starting voltage that you go to, but you’ve got to go to maybe two to three times the starting voltage to make it strike. Then it strikes with some regularity..
HT: .. We’re – I see we’re getting near the end of the tape, but I want to make a comment here for the tape because .. I think inadvertently you answered a very important question of mine when you talked about the .. coming back to NCR and your previous work after the involvement in the Navy classified work. And your statement about having to get back into computers all over again because there had been no connection. And I think that answers the – the main reason that I’ve been probing at that area. Not because I’m interested in the classified material, not because I’m interested in knowing what people were doing, what I was
[Page 164]
interested in was the by-product of that work into the computer field and in your case, you’ve answered the question. The by-product was zero.
JD: Well, it was for the Navy, but it was not for OSRD.
HT: Oh, no, no, no, no. I’m talking about that – that one project of yours.
JD: That one hush-hush one.
HT: Right, and I’m not sure but that the project I mentioned in England didn’t have similar .. neglgible, or at least not very large impact on later developments.
JD: Well, when you’re solving a very high priority, urgent problem for the military, you don’t put much stock in whether or not it’s going to be very profitable to you afterwards or not.
HT: Right.
JD: Oh, the aircraft – aerospace industry may, but we didn’t. And we couldn’t.
We couldn’t orient our work to do that.
HT: Mhm.
JD: We had to accomplish what the Navy wanted,
HT: Right.
JD: or any other branch of the Armed Services.
RM: That was a special purpose – special purpose device.
[Page 165]
JD: That was a very special purpose
HT: Right.
JD: as you can see in reading my notebooks.
HT: Right. And as I say without – without going into any of the detail, even the most minute detail, you’ve essentially answered a very important question
JD: All right.
HT: which was the only reason that I was interested in – in that topic in tracing the development of ideas and the impact that people had on each other
JD: Yes.
HT: and the exchange of ideas in projects they were working on.
JD: Well, I couldn’t have, in good conscience, hahave done otherwise and tried tc use all that stuff from NDRC in trying to solve a problem for the Navy when it had no connection whatever.
HT: Mhm.
JD: And if I had tried to bend it around like that, .. I wouldn’t have been doing my duty, I can tell you right now.
HT: Mhm.
JD: I couldn’t do it.
[Page 166]
HT: Well, since we’re running near the end of our time, what I’m going to do is turn the tape off and then I’d like to beg a few more minutes of
your time.
END OF TAPE 2, Side 2
End of the transcript of January 17, 1973