Deborah Greer: Okay. Welcome. My name is Deborah Greer, and I’d like to wel­come you here today for the Women of the ENIAC pre­sen­ta­tion. When I first read about these women, it was in the Wall Street Journal arti­cles by Tom Petzinger [1], and I was just real­ly fas­ci­nat­ed by their sto­ry. So I looked a lit­tle fur­ther and I found anoth­er arti­cle by W. Barkley Fritz which was also very good. And then I got real­ly inter­est­ed in their sto­ry and I just kept think­ing it’d be so great to meet them and to hear their sto­ries per­son­al­ly. And then I thought well I bet there’s a lot of oth­er peo­ple at Microsoft that would also like to meet and hear them. So for­tu­nate­ly I work for some­one who also thought this was an excit­ing and worth­while project and gave me to the go-ahead to make it happen.

Before I intro­duce my man­ag­er, though, I just want to tell a lit­tle bit about the for­mat. Dennis is going to come up and give a brief intro­duc­tion, and then Jean and Kay are going to each give a short bio. And then Maureen will be pos­ing ques­tions as our mod­er­a­tor today to take them through a lot of the impor­tant ques­tions that all of us want to hear answers to. And then at the end there should be time for audi­ence ques­tions as well.

Now I’d like to intro­duce you to the Director of Product Developement Resources, Dennis Schnabel.

Dennis Schnabel: It’s indeed an hon­or to be able to intro­duce our guests today two of the orig­i­nal pro­gram­mers from the ENIAC com­put­er in 1946. It’s 50 years, and it’s real­ly mind­bog­gling some­times to think about the changes that’ve hap­pened in the com­put­er indus­try over the last 50 years. Even in the 25 years that I’ve been in the com­put­er indus­try it’s real­ly incred­i­ble. I remem­ber the main­frame com­put­er I used back in grad­u­ate school filled up prob­a­bly about a third of the size of this room and had spe­cial air con­di­tion­ing and had less pow­er and capac­i­ty than the work­sta­tion I have on my desk­top right now. So it’s real­ly amaz­ing to see how much change has hap­pened. But for many of us 50 years seems like a real­ly long time, but real­ly in the span of any of the engi­neer­ing pro­fes­sions it’s real­ly a short peri­od of time. And I thought I’d give you a cou­ple of exam­ple of that.

Civil engi­neer­ing, for exam­ple, goes back thou­sands of years to the Romans build­ing of the bridges and roads, and the Chinese build­ing the Great Wall of China. And mechan­i­cal engi­neer­ing, the first pis­ton engine was built in 1690 and it took anoth­er 79 years for James Watt to build the first steam engine in 1769. Another 61 years for the first rail­road to be built in 1830. And anoth­er 55 years for the first car to be built by Karl Benz in 1885.

Even elec­tri­cal engi­neer­ing, which is rel­a­tive­ly young, also has had a long times­pan, real­ly. The first con­denser, the Leiden jar, was built in 1745 and it took 92 years for the first tele­graph to be built in 1837. And anoth­er 40 years for the first elec­tric light to be built in 1877. So you can see that [the] pro­gram­ming pro­fes­sion, soft­ware engi­neer­ing, is real­ly a very young pro­fes­sion and we still have got a lot to learn, a tremen­dous amount to learn. 

But we owe a tremen­dous amount of debt to the two women who are with us today because they were one of the first pio­neers of our orig­i­nal pro­fes­sion. When they start­ed work­ing on the ENIAC, it was clas­si­fied, so they did­n’t have access to the machine until their secu­ri­ty clear­ances came through, so the only thing they had was the log­ic dia­grams to try to fig­ure out how to make this machine work. And they had to recon­fig­ure the machine for every pro­gram. So they had to reshuf­fle cables, reset oper­a­tional switch­es and all sorts of things. That’s real machine-level pro­gram­ming. A year lat­er in 1947, the ENIAC was con­vert­ed to the first stored-program com­put­er; a great deal of work by Jean to make that happen. 

But it was inter­est­ing to see that many of the same sorts of prob­lems they dealt with, we still seem to deal with our­selves. For exam­ple, how to make the code run as fast and as effi­cient as pos­si­ble. How to reuse the code as effec­tive­ly as pos­si­ble. I guess some of these prob­lems just nev­er seem to go away.

So with that, I think it’s an incred­i­ble hon­or on my part to wel­come these two women to Microsoft. Two of our orig­i­nal pio­neers, and we have a great deal of debt owed to them. And I’m real­ly look­ing for­ward to hear­ing more about what they have to say. So wel­come Jean Bartik and Kay Mauchly Antonelli. Thank you.

Maureen: Going to start off with some short bios of where they saw their careers and what they saw their con­tri­bu­tions can be. Do you want to start?

Jean Bartik: I’ll be hap­py to, yeah. Thank you so much for being here. You can imag­ine how dif­fer­ent it is today see­ing your cam­pus. As a mat­ter of fact, I don’t know if any of you know this, but I would say it looks like a very pros­per­ous com­pa­ny. And cer­tain­ly your work­ing spaces… Let me tell you, if I’d had any­thing like that I can’t imag­ine what I’d have been able to do. And cer­tain­ly I want to say that I did not choose this career. I just felt that I was lucky in being in the right place at the right time. How could I have cho­sen a career that I did­n’t even know exist­ed? And as a mat­ter of fact it did­n’t exist until the ENIAC came along.

So any­way, I’m just incred­i­bly grate­ful­ly that I hap­pened to be there at that time, for many rea­sons, but the biggest one that it was a lot of fun and it cer­tain­ly affect­ed the rest of my life in a very sig­nif­i­cant way. But I start­ed out as a farm girl from Missouri. I’m the sixth of sev­en chil­dren and the third daugh­ter, so my moth­er had a cook and she had a house­keep­er before I came along. So being the third girl, I actu­al­ly worked in the field and plowed corn and rode hors­es and put up hay and did all this kind of stuff. But at the same time, as women have for­ev­er, I had to help with the house­keep­ing. [inaudi­ble sentence]

So any­way, I went off to col­lege at the age of 16 in 1941, and when I start­ed I thought, Well, I’ll be a jour­nal­ism major,” and the rea­son I did, even though I was very good in math, I could­n’t think of any­thing to with math except teach school, and I did­n’t want to teach school. And I’m sure that you can’t imag­ine how iso­lat­ed we were in com­par­i­son with how you are today. I read about all these excit­ing places, but I’d nev­er been any­where or seen any of it. So any­way, I went off to col­lege in 1941. Of course Pearl Harbor was bombed, and imme­di­ate­ly all the men were swept away from the cam­pus, and in the spring of 42 it almost like a girl’s school because all the men were gone. 

But then the next year, they brought in V‑2 and V‑5 pro­grams, and the V‑2 peo­ple were men that had been at sea, and we real­ly thought it was real­ly some­thing. There were sailors that came in that had been out and seen the world. So any­way, I took my math cours­es, trigonom­e­try, ana­lyt­ic geom­e­try, and physics, with these sailors. Now, you have to under­stand that I was the only math major in the school at this point, not just the only woman math major, the only math major in that school.

So any­way, the cours­es then that were giv­en afterward…I did have a guy from South America who was there going to school that took advanced cal­cu­lus and some of these cours­es. But I had cours­es all by myself because the head of the depart­ment said, Well, we offer the degree, and if we offer the degree we have to offer the cours­es and there­fore we just…give em,” so that’s what they did. I had cours­es all by myself. But any­way, when it was time to… I fin­ished my course­work at the end of 1944, so I was busi­ly look­ing for a job and I did­n’t want to teach school, and I was absolute­ly inun­dat­ed because they were so des­per­ate for math teach­ers in all the schools around.

But one of my teach­ers came in one day and said to me, I just got this let­ter from math soci­ety and they’re look­ing for math majors at the University of Pennsylvania, but they’re doing cal­cu­la­tions for Aberdeen Proving Grounds.” So she said to me… She used to work for Wright-Patterson Air Force Base and they had a dif­fer­en­tial ana­lyz­er. So there were three in the United States at that time, one at MIT, one at Wright-Patterson, and there was one at Penn. And she said, Oh you must go to Penn because they’ve got a dif­fer­en­tial ana­lyz­er.” Well, I had no idea what a dif­fer­en­tial ana­lyz­er was, but I knew one thing: Pennsylvania was not Missouri. 

So any­way, I applied for the job and I did­n’t get an answer, and my father was a school­teacher and he used to come home every day telling me about anoth­er math job and he was real­ly putting the pres­sure on me to get to work. So any­way, final­ly they sent me a telegram, said I was hired and to get there as quick as pos­si­ble. Well, I was on the mid­night Wabash the next night, out of Missouri. 

So I got there and they were cal­cu­lat­ing tra­jec­to­ries to make fir­ing tables for the new guns that they were devel­op­ing. And they had about 60…they called us com­put­ers. I was hired for the grand total of $2000 a year with a $400 bonus for work­ing Saturdays. So I took this [?] ride to Philadelphia for $2400 a year.

Maureen: We don’t get a bonus for work­ing Saturdays, by the way.

Bartik: Oh, you don’t?

There were about 60 peo­ple that were cal­cu­lat­ing these tra­jec­to­ries using Monroe and Marchant cal­cu­la­tors. I learned how to do it and it was fun, and Philadelphia was a lot of fun. But about in May I guess it was, they announced… Well, they just sent a notice around that any­body that would want to work on this new machine called the ENIAC, any­body could apply. But I nev­er had any idea what it was, but I knew it was­n’t push­ing a Monroe and Marchant cal­cu­la­tor. So, you know…career planning. 

So any­way, I applied and went over and they just talked to us a lit­tle bit. We nev­er saw the machine or any­thing. So then they called us in and Herman Goldstine, who was the Army offi­cer liai­son com­ing in from Aberdeen, inter­viewed me. So Herman said to me, What do you think of electricity?”

So I said, Well, I had a physics course and I knew that E=IR.”

So he said, No, I don’t mean that. I don’t care about that. Are you afraid of it?”

I said, Well, no I was­n’t afraid of it.”

So then he explained to me the rea­son he’d asked me the ques­tion was that the ENIAC had dig­it trays and data trays and switch­es and cables to plug in, so he want­ed to know that I would­n’t be afraid to set switch­es and plug in cables.

So any­way, that’s my…well, wait a minute. They picked five peo­ple, which I did­n’t real­ize I was the only one in the group that was final­ly picked that was ever inter­viewed. All the rest of them were picked for var­i­ous oth­er rea­sons. Anyway, there were five picked and I was the sec­ond alter­nate. So I thought oh, well, that’s it. So time passed and on Friday after­noon a man­ag­er called me in and said, Can you be ready to go to Aberdeen on Monday?” because they were send­ing the ENIAC pro­gram­mers (or what­ev­er we were at that time; we were called com­put­ers) to Aberdeen to learn punch card equip­ment because the I/O of the ENIAC was punch cards and our print­er was an IBM tabulator.

So I said, Sure.” And it turned out that the fifth girl that was select­ed had a very nice apart­ment in Philadelphia and hous­ing was very hard and Aberdeen was a hell­hole, so she decid­ed she did­n’t want to go so she refused. Okay, so the first alter­nate was away on vaca­tion. So they called her up and said she had to come home and go to Aberdeen. But she also knew Aberdeen was a hell­hole, so she did­n’t come home.

So I end­ed up and so much for career plan­ning. That’s how it happened.

Maureen: I think many of us have had our career plan­ning go upon a sim­i­lar path. Kay, how did you get involved?

Kathleen Antonelli: Well, it cer­tain­ly was­n’t as inter­est­ing as Betty Jean’s. I had gone to a girl’s school all my life, and I went to a girl’s col­lege. I was a senior in col­lege when the war broke out in 1941, and I grad­u­at­ed in 42 with a major in math­e­mat­ics, with no idea what in the world I was going to do with it. I just loved math­e­mat­ics, I loved physics, loved every­thing like that, but I did­n’t want to teach.

And there I was with my degree wav­ing in my hand but not know­ing what to do with it. Within a week after I had grad­u­at­ed, there appeared an ad in the Philadelphia Bulletin say­ing Wanted: women col­lege grad­u­ate math majors.” Moore School, and that was the build­ing of the ENIAC, and it was clas­si­fied. We were not told any of the infor­ma­tion about it. But we knew that it was going to be a machine that would do these same tra­jec­to­ries maybe in twen­ty sec­onds, in com­par­i­son with what it had been. But we were nev­er allowed to see it or any­thing like that, we just knew that it was a project that was going on at the Moore School.

Finally, when we had VE day, which was like May or June of 1945, our job there at the University of Pennsylvania was com­ing to an end. The war was end­ing. They knew the war in Japan would end very soon, and would we need any more fir­ing tables? Well, they decid­ed that since the ENIAC also was about to come online, it was impor­tant that they train some women—women, because the men weren’t yet back from the war. That was the whole thing. The men weren’t back from the war. It was almost a hun­dred women they had work­ing there in Philadelphia, and anoth­er hun­dred in Aberdeen Proving Grounds just hand-computing these tra­jec­to­ries. And here was this ENIAC about to come online. The war was over, it had been designed with the idea that it was going to solve all our prob­lems, and yet the war was over and it was­n’t even run­ning. But they sent us to Aberdeen Proving Grounds, and Betty and I and for oth­er girls—three oth­er girls at that time, went to Aberdeen to learn all about IBM equip­ment because the IO was on IBM machines. 

And when we came back after ten weeks of learn­ing all about IBM equip­ment, then we were told that we had to learn what we could so that we would be able to pro­gram the ENIAC. What for? To do tra­jec­to­ries. Well, why did they want tra­jec­to­ries? Well, that was what it was built for so that’s what we were going to do.

So, no man­u­als had been writ­ten. None what­so­ev­er. Some of the engi­neers who worked on the project gave us their great big blue­prints or wiring dia­grams, and their block dia­grams and things, and we worked our way through these things to find out what every tube would do. Now, the basic ele­ment of the ENIAC was some­thing called an accu­mu­la­tor, which had 500 tubes in it. So essen­tial­ly these were like ten decade coun­ters in each one, and once you learned what one tube did, you could fig­ure out what the oth­er ones were sup­posed to do.

Well, there were very many dif­fer­ent parts of the ENIAC. If any of you have seen pic­tures of it, you know it was eighty feet long and was made up of forty pan­els, each one two feet wide. And twen­ty of them were called accu­mu­la­tors, and they were to do addi­tions and also to store num­bers. The com­plete mem­o­ry of ENIAC itself was 200 dig­its. It oper­at­ed at 5,000 add times in one sec­ond, which then was the fast­ed thing in the world, incred­i­bly faster. 5,000 times faster than any oth­er machine oper­at­ing. It seems like noth­ing today. The full pow­er of that ENIAC, all eighty feet of it, could fit on a lit­tle chip one inch square nowadays.

Eventually, we taught our­selves how to pro­gram a tra­jec­to­ry for the ENIAC. And then, as long as I was there, nev­er once did we actu­al­ly run a set of tra­jec­to­ries, although that was what it was made for.

Maureen: So was your respon­si­bil­i­ty to…you evolved your­selves into the orig­i­nal programmers?

Antonelli: Yes. Our respon­si­bil­i­ty was to decide just exact­ly how we would fit all the math­e­mat­ics of a tra­jec­to­ry onto the ENIAC.

Maureen: What oth­er appli­ca­tions did you have to work on?

Antonelli: Other applications?

Maureen: Well, you said that you did­n’t do tra­jec­to­ries, ever.

Antonelli: That’s right. Well, the very very first prob­lem, the very first day that the ENIAC was final­ly declared it is now ready to work, although it had­n’t been ful­ly test­ed, our test will be the fea­si­bil­i­ty of the H‑bomb. It was a tremen­dous, tremen­dous prob­lem that had been brought in from Los Alamos. It was Dr. Fermi and Dr. Teller’s prob­lem, and it was the fea­si­bil­i­ty of the trig­ger which they had designed.

So we put that prob­lem on the ENIAC. Now, some­one had asked us a lit­tle while ago about the tubes and the reli­a­bil­i­ty of them. There were many tubes that failed, but the ENIAC was built in such a way that a tube did not have to gen­er­ate a cer­tain amount of pow­er— This is a vac­u­um tube I’m talk­ing about, which in those days were pret­ty large. They were about five inch­es tall, each one, and there were 18,000 in the ENIAC. Which gen­er­at­ed a tremen­dous amount of heat and every­thing. These tubes were not oper­at­ed full pow­er. Any amount of pow­er would turn them on. So they only oper­at­ed in an on or an off posi­tion. If they were on at all, they trans­mit­ted the pow­er. If they were off, they did­n’t. And that was the basic thing.

So when we first start­ed to oper­ate the machine, we found out that there were a num­ber of tubes that had actu­al­ly failed, and anoth­er bad sol­der joints that had gone on and so on. Once those things got ironed out, the machine was real­ly quite reli­able and it oper­at­ed very well. So the prob­lem that came from Los Alamos at that time, we did not pro­gram it. The physi­cists who came from Los Alamos had actu­al­ly pro­grammed it. They had stud­ied all about the ENIAC because they had Top Secret clear­ance, and they were able to have access to these wiring dia­grams before we did, even. So they were able to pro­gram it and we helped them, all of us girls (girls we were called in those days) helped them pro­gram it and help them run it. It took about two and a half months.

I don’t know what it proved about the hydro­gen bomb, whether the answer was yes or no, the trig­ger was fea­si­ble or not, I have no idea about that. But what we did find out was that the ENIAC actu­al­ly worked, was actu­al­ly reli­able, actu­al­ly did every­thing that was planned, and it looked like no mat­ter what prob­lem you put on there, the ENIAC would even­tu­al­ly be able to solve it. So on the 15th of February of 1946, they had a huge demon­stra­tion and announced to the world, yes that this machine worked.

Maureen: So the first fea­si­bil­i­ty study on computers.

Antonelli: Yes, yes.

Maureen: The first fea­si­bil­i­ty val­i­da­tion. Jean, how did you feel at the time of what you were work­ing on, and how much did you have to under­stand what was being done, the appli­ca­tion itself?

Bartik: Well, the ENIAC we knew because we’d all cal­cu­lat­ed it over and over. I mean, the tra­jec­to­ries we knew, because we all cal­cu­lat­ed them over and over again. So that was­n’t a prob­lem. Our prob­lem from the very begin­ning was find­ing out how to fit this prob­lem on a machine with these switch­es. We had to learn how to reuse code. They had some­thing called a mas­ter pro­gram­mer so that we could iter­ate a pro­gram any num­ber of times and then after we’d done it we could go do some­thing else. We also could check the sign of a num­ber, whether it was plus or minus, so we could arrange that…that would be the if/then state­ment or what­ev­er it is.

Anyway, let me go back, though, to some­thing that you brought up before about the reli­a­bil­i­ty. When Pres and John first pro­posed the ENIAC, all of the pro­fes­sors at Penn and Moore School, and every­one else they talked to said they were nuts to think that they could use vac­u­um tubes and have them reli­able enough to build a com­put­er out of. Well, Pres Eckert was all of 24 years old at that time, and a hot-shot engi­neer. Well, Pres Eckert said, I don’t believe it. I can do it.”

Now, how could he do it? That was by run­ning the vac­u­um tubes at 50% of their capac­i­ty so that they did not have to work very well. And as she said, they ran them as either they were off or on. So he designed it in such a way that if 50% of the tubes weren’t work­ing at full capac­i­ty, then the machine still worked. 

Now, the tra­jec­to­ry prob­lem was a very very impor­tant prob­lem for us. First of all, it was the accep­tance test of the ENIAC. The sec­ond thing was it trained us to pro­gram the ENIAC. We had two of the women that were part of our team to cal­cu­late a tra­jec­to­ry exact­ly the way the ENIAC did so that it was a test to debug it, because we knew exact­ly, at every stage of the cal­cu­la­tion, what the num­ber be every­where. So we could diag­nose it down to a vac­u­um tube.

Maureen: So you would actu­al­ly break it up into sec­tions that you would man­u­al­ly compute.

Bartik: Well, if the [com­put­er?] hav­ing prob­lems. Now, the stan­dard tra­jec­to­ry, the way we assumed that our answers were cor­rect, before we ran the real prob­lem, we ran the test tra­jec­to­ry. And if it were cor­rect, then we ran the real prob­lem. The we ran the test tra­jec­to­ry again. Now, if both test tra­jec­to­ries turned out cor­rect, then we assumed the one in between was cor­rect. And that worked very very well. So it was quite a while, I would say, before the pro­fes­sors at Penn and every­where real­ly believed that the ENIAC could work and would give worth­while answers. And it cer­tain­ly did because it was in use for ten years, and Barkley Fritz has enu­mer­at­ed 100 sci­en­tif­ic prob­lems that it actu­al­ly cal­cu­lat­ed. So it did work. It was not just a pub­lic­i­ty stunt or any­thing like that.

Maureen: As pro­gram­mers, what was the most inter­est­ing pro­gram you had to do on the com­put­er past the feasibility?

Antonelli: The most inter­est­ing pro­gram? Well, for me, short­ly after the ENIAC was announced, the uni­ver­si­ty began to get calls from all over the world. People writ­ing, they’d heard that there was going to be this won­der­ful machine. And before that, pri­or to that, when all com­put­ing was done either by hand or with a desk cal­cu­la­tor, peo­ple had prob­lems that they would like to solve. They called them hun­dred year prob­lems. They would take so long to do that they knew they’d nev­er be able to fin­ish them in their life­time. And all of a sud­den, here was a machine that it was a hun­dred year prob­lem and the thing went 5,000 times as fast as any oth­er machine. You’d be able to do this in less than a year, obviously. 

And so peo­ple who had pet prob­lems of some kind decid­ed to write to Penn to see if they could be priv­i­leged to oper­ate. So one of the prob­lems, and the one that is most mem­o­rable to me, was brought there by a Professor Douglas Hartree from the University of Cambridge in England. He has a pet prob­lem that he had been work­ing on for years, and it had to do with the air flow around the wings of an air­plane or some­thing like that. So he had devised some won­der­ful cal­cu­la­tions, and he had been doing these by hand. He said when­ev­er he trav­eled on the train or any­thing like that, he’d always get out the note­book and keep work­ing on this com­press­ible bound­ary lay­er. So when he heard about that, he wrote to the Army, I sup­posed, and asked per­mis­sion to come. So he came with his prob­lem. I was assigned to work with him and to help him to code his prob­lem. And that was most mem­o­rable because he was a won­der­ful, metic­u­lous kind of per­son. He knew an awful lot about com­put­ing, and in those days peo­ple did­n’t know so much about com­put­ing. We did months of study just to find out how round-up error built up and var­i­ous math­e­mat­i­cal prob­lems that became involved in that. 

Oh, I had one final note on that. In November of this past year, Aberdeen Proving Grounds had a big cel­e­bra­tion to cel­e­brate the 50th anniver­sary of acquir­ing the ENIAC. And lo and behold, they have since the ENIAC acquired a num­ber of faster and faster and faster machines. And do you know that they are still work­ing on that com­press­ible bound­ary lay­er problem?

Maureen: Talk about job security. 

Bartik: I’d like to add a note about Dr. Hartree. When you’re at the begin­ning of an indus­try like this, peo­ple talk about all kinds of things. And one of the things they talked about was what kind of peo­ple you need­ed to run these com­put­ers, because at this time, all these prob­lems were sci­en­tif­ic. So peo­ple said, Well, we need a numer­i­cal analyst.”

So we would say, What’s a numer­i­cal analyst?” 

And they would say, Well, we don’t know what a numer­i­cal ana­lyst is, but Douglas Hartree is one.”

Maureen: Jean, what was the most inter­est­ing appli­ca­tion you had to work on?

Bartik: Let me go back to say how they ran it at that time. What they did, you would have a sci­en­tist who had a prob­lem to do, and then a pro­gram­mer would be assigned to the sci­en­tist to the problem. 

To me…Betty Holbrook and I pro­grammed a tra­jec­to­ry that was run for the demon­stra­tion and that was by far the most excit­ing. And I’ll tell you why it was excit­ing. People came from all over the world [to] see this, and this program…well, we worked night and day, and I’m sure many of you peo­ple have done the same thing. But because they said two weeks ahead of time, Can you have it ready?” and we said, Sure. No prob­lem.

Maureen: Did you have some­one named Steve Ballmer work­ing with you? This is your inter­view, I’m sor­ry. Just could­n’t resist.

Bartik: Anyway, so we worked night and day. And right before, it worked fine except it did­n’t stop when it hit the ground, it dug a lit­tle hole. And I’m sure all of you have had a hard time stop­ping a do loop at the right time. So what we did was to run the tra­jec­to­ry and we took the punch cards to the tab­u­la­tor and read it out, ripped the sheet off and gave it to them as demon­stra­tion. They took it home with them. And then this pro­gram actu­al­ly ran faster than it took the bul­let to trace the trajectory. 

And the oth­er thing was, and you [Antonelli] should tell about this because I don’t know this sto­ry that well, but the ENIAC lights flashed up and down because you could see the num­bers build­ing up in accu­mu­la­tors and going down—

Antonelli: Oh, let me tell that story.

Bartik: So you tell it.

Antonelli: Well, the ENIAC not only did all its cal­cu­la­tions behind the scenes because all the vac­u­um tubes were stick­ing out the back, but if a tube was acti­vat­ed at all, there was a lit­tle neon light in the front of the ENIAC, and it showed exact­ly what posi­tion the—there was what they called a…what’d they call it?

Bartik: Decade counter.

Antonelli: A decade counter, which meant there were ten bulbs, ten vac­u­um tubes in the back. But every time one of these vac­u­um tubes lit up, this lit­tle neon light in the front was sup­posed to—did actu­al­ly light up, and that was how we test­ed it, by see­ing when the lit­tle neon lights lit up.

Well, the pho­tog­ra­phers came (this was about two weeks before the big dis­play) and they want­ed to take a movie of the ENIAC work­ing. And when they took this movie, and in those days we only had black and white, and the neon lights just showed up black. So you could­n’t see a thing. And we said how in the world are they ever going to do that? So the sto­ry that the news­pa­pers final­ly picked up was that they went out and they put ping pong balls on there. Because when­ev­er you see a pic­ture of the ENIAC, you will see for each num­ber on the decade counter, a great big round ball. Well, actu­al­ly, what actu­al­ly hap­pened was that John Mauchly him­self had gone out and bought a case of lit­tle white bulbs, and they spent two nights, just every­body, all the engi­neers, paint­ing 0 up to 9 on 2,000 bulbs. And then they unscrewed the lit­tle neons and screwed in these lit­tle real lights. It was only for that demon­stra­tion, but when­ev­er you see a pic­ture of it, you will see oh, they actu­al­ly paint­ed the num­ber 1 to 9 on these lit­tle lights. So when­ev­er you see it, you’ll see this demon­stra­tion thing. 

Maureen: And we thought we invent­ed marketing.

Bartik: Well, wait a minute.

Maureen: Keep going. Please.

Bartik: Back to the prob­lem. So that was very excit­ing and I did write work on a prob­lem with [tab?], it was called the [tab?] prob­lem. But real­ly the most excit­ing pro­gram that I worked on was when we turned it into a stored-program com­put­er, because very ear­ly on it was pret­ty obvi­ous that… I mean, you weren’t going to get very many pro­grams to run on the ENIAC because not only did you have this prob­lem with pro­gram­ming them, but you had of set­ting them up and test­ing them all the time. So—

Antonelli: I think we have to explain that the ENIAC was not a ser­i­al machine, it was a par­al­lel machine, and you could have three or four oper­a­tions going on at the same time. And dif­fer­ent oper­a­tions took a dif­fer­ent length of time. Like, an addi­tion, when you add one num­ber to anoth­er, just took what they called one add time. A mul­ti­pli­ca­tion of a 10-digit num­ber by a sec­ond 10-digit num­ber took 14 add times. A divi­sion, or a square root, you could­n’t deter­mine how long it was going to take. It depend­ed on the num­bers that were involved.

So if you were try­ing to run a par­al­lel oper­a­tion and you were doing a divi­sion here and a cou­ple addi­tions over here and a mul­ti­pli­ca­tion over here, you had to make your pro­gram so that the thing that was going to take the longest length of time was the one that start­ed up the rest of the pro­gram. And you would have to be very care­ful to do this and to bal­ance all your work. So even­tu­al­ly it became ridicu­lous that you were spend­ing all your time count­ing puls­es when each oper­a­tion would end instead of just doing a ser­i­al oper­a­tion. So what Betty it talk­ing about is that even­tu­al­ly, with­in the year, ENIAC was…not redesigned, but extra pieces were put on it so that it was able to oper­ate seri­al­ly and in what you might call a stored-program function.

Bartik: What hap­pened was the drag func­tion for the shelves was put on func­tion tables, and it was built with three func­tion tables with a hun­dred entries. So once we did­n’t do tra­jec­to­ries, you did­n’t need these func­tion tables for a drag func­tion. So I think Pres and John had thought of this ahead of time, but in any case you could use this func­tion table to store pro­grams. And what we did was to design an instruc­tion set, and it was designed a sim­ple accu­mu­la­tor. So the device that they had made was to be able to pull two dig­its at a time out of the func­tion table, because the func­tion table had twelve digits…

Antonelli: Yeah, twelve dig­its across, a hun­dred deep.

Bartik: So you could pull two at a time, and we had some­thing called a mas­ter pro­gram­mer, which was the con­trol thing of the ENIAC. You fed it into the mas­ter pro­gram­mer and it decod­ed it into a pro­gram out­put pulse which you could use to stim­u­late an oper­a­tion and then you’d go back and get the next one, etc.

So then, this instruc­tion set looked very much like the instruc­tion set of mod­ern com­put­ers. You had add, sub­tract, mul­ti­ply, and divide, but it all worked around a cen­tral accu­mu­la­tor. And this slowed the machine down tremen­dous­ly. But, it was much eas­i­er to pro­gram and you did­n’t have to keep reset­ting these switch­es and dig­it trays so the prob­a­bil­i­ty of errors went down and every thing else. So you could change from one pro­gram to anoth­er very eas­i­ly just by chang­ing a bunch of switch­es on this func­tion table.

We were the only gen­er­a­tion that used the ENIAC in its orig­i­nal form. Because every­body after that, after it was moved to Aberdeen, used it as a stored-program com­put­er. And the dif­fi­cul­ty with turn­ing it into a stored-program com­put­er was to be able to get all the stuff on it, and also to get as min­i­mum an instruc­tion set so we could get it all on, but get enough so we could do all the prob­lems we want­ed to do. So that was a lot of fun to do that.

Maureen: How did your job change and your career evolve with your asso­ci­a­tion with the ENIAC?

Bartik: See, they’d already designed and thought of EDVAC before. So EDVAC, which was the stored-program com­put­er that they were build­ing for Aberdeen, was already in the works. And also Pres and John had this dif­fi­cul­ty with patents with Penn so they left and formed their own com­pa­ny. So I left to go work for Pres and John. And then they were busi­ly build­ing the first com­mer­cial com­put­er, which was the UNIVAC, and as many of you may remem­ber, at one time in this indus­try a UNIVAC meant com­put­er” the same way Frigidaire means refrig­er­a­tor.” You know, you’d say, Well, do you have a UNIVAC?”

That was one of the most excit­ing times of my life because first of all, I turned from a pro­gram­mer into a log­i­cal design­er because I did log­ic design on UNIVAC I and put in the check cir­cuits and stuff like that. And the team…and I’m sure you’ve seen these team efforts here, but the team that designed UNIVAC I was like twelve of us. So it was unbe­liev­able esprit de corps and we worked…all the time. It’s about all you do. But, I loved it. I real­ly loved it. 

Then one of the things that I think peo­ple who have nev­er been there can’t under­stand how much fun it is to work your mind and your body at full capac­i­ty for peri­ods of time. It’s very exhil­a­rat­ing. And then of course the sense of accom­plish­ment once it’s done is incredible.

Maureen: That relates entire­ly to the work that we do. That has­n’t changed. [To Kay Antonelli:] How did your job evolve, and your career evolve?

Antonelli: Well, I went to Aberdeen… When the ENIAC was about a year old, it was moved to Aberdeen Proving Grounds, where they had build a whole build­ing to house it. Then they moved the machine down. They just tore down a wall in the side of Moore School and took the machine out. They had built a boat in the base­ment and had to get it out. So they tore down a wall and took it out, and when it got the Aberdeen, of course every­thing had to be retest­ed. The whole machine had to be reassem­bled and retest­ed. So the first year I was down there, about three quar­ters of the year, any­way, was spent just reassem­bling the ENIAC and get­ting it all test­ed out.

Then after John Mauchly had been the chief engi­neer— No, not the chief engi­neer, but he was the design engi­neer and inven­tor of the ENIAC, and was very much inter­est­ed in what it was doing once it got to Aberdeen, and he kept com­ing down to vis­it it, and even­tu­al­ly me. And then the fol­low­ing year, in 1948, we got mar­ried and then I left the busi­ness entire­ly. Betty and Betty Holberton went to work for John Mauchly and Pres Eckert, but I was need­ed at home, at his home, and so I nev­er went back to work.

Maureen: What do you think was your per­son­al con­tri­bu­tion? When you look back and think about what you did, what was the spark that you think you brought to the table?

Antonelli: Well, I don’t know. I thought at the time that I had a lot to do with how we actu­al­ly did the design, the pro­gram­ming, for how ENIAC itself would be pro­grammed. That was before it was changed to a machine that…the stored-program.

Bartik: Well, I’d like to say what I think she brought to it. Betty Holberton and I were the work­hors­es of the crew, and very log­i­cal, A B C D and all this kind of stuff. Kay is much more cre­ative than I will ever be, and she was the one that helped us real­ly under­stand this busi­ness about reusing code. How to split it up, reuse code, and all this. She was very cre­ative and I real­ly think she was the most cre­ative per­son on [inaudi­ble; crosstalk]

Antonelli: [crosstalk] Wow, thank you Jean. Oh, my.

Maureen: Kay, what did Jean bring to the table? Give you some equal time. What did Jean bring to the table?

Bartik: Oh, please.

Antonelli: I think, num­ber one her great good humor. We had long tedious days and night. We were sup­posed to work for eight hours. That eight hours went into ten, eleven hours most nights.

Maureen: And that was with­out Jolt cola.

Antonelli: That was what?

Maureen: That was with­out high-caffeine colas.

Antonelli: Oh, no. No caf­feine, no Coke or any­thing like that. As a mat­ter of fact, I don’t think we ever brought any­thing to eat into the ENIAC room, did we?

Bartik: No.

Antonelli: No. But, num­ber one was, as I say, the high spir­its which you brought to the whole group. And I think she also was very very cre­ative, and her sub­se­quent career proved all of that.

Maureen: [To Jean Bartik:] What do you look back, and what do you see as your shin­ing star? What are you most proud of?

Bartik: Jeez…

Maureen: I know. It’s a trick ques­tion. We did­n’t have that on the list that we went over.

Bartik: I…tell you the truth, I nev­er even think about that.

Maureen: Okay.

Bartik: I mean, I was there, I did the job. I have no delu­sions that if I had­n’t have been there some­body else would­n’t have done it. I mean, I did not… I think I was full of it, and would’ve done some­thing because there’s no doubt about it that I was ready for an adven­ture and some­thing that was new and dif­fer­ent and excit­ing. So I have no doubt if I had­n’t have been picked for an ENIAC pro­gram­mer I’d’ve been off doing some­thing else.

Maureen: First woman on the moon. What do you think?

Antoneli: I think in gen­er­al that most of us just thought it was part of our job. I mean, we just…this is some­thing that had to be done and we went right ahead and did it the best we were able to, with­out any­body to show us how to do it.

Maureen: Did you have any sense that you were cre­at­ing his­to­ry, creating—

Bartik: Oh my God, we talked about it night and day. I mean, of course. Socially, that’s in fact my ex-husband, or my hus­band at the time, got very bored hear­ing us talk about it because socially—

Maureen: This is my life.

Bartik: Right. We would go into a room and we’d start out with the pleas­antries of life, and the next thing you know we’d be into talk­ing about com­put­ers and where they were going to go and what we were going to do and all that kind of stuff. I mean, we thought about it night and day.

Antonelli: Yeah. And you know, look­ing back it did­n’t seem as if there was a lot of romance going on at the time. We were each doing our own, and yet Betty mar­ried one of the engi­neers at the Moore School. Fran Bilas mar­ried the Army engi­neer that was assigned to the ENIAC. I mar­ried John Mauchly. Betty Snyder mar­ried John Holberton, who was our imme­di­ate super­vi­sor. And one of the girls mar­ried a dentist.

Maureen: Okay. I want to get into some of the peo­ple that you’ve been talk­ing about and get your impres­sions on some of them. A lot of these are very sig­nif­i­cant names to us in the indus­try, peo­ple that we’ve kind of come to admire. But first, you brought us some pic­tures, and I’d like to show the pic­tures, and if you could tell us who they are. This one is kind of easy to talk about.

Bartik: When I first came to Philadelphia, I went down to have my pic­ture tak­en. And there was a lit­tle old man down on 4th Street, 4th and Market in Philadelphia, and he real­ly took a shine to me. So I had dozens of these pic­tures that he took of me, and he used to shine them on the wall on 4th Street so the trol­ley could see what mag­nif­i­cent pic­tures he was tak­ing. And I was 20 years old that time.

And this is Betty Snyder Holberton and me on my wed­ding day. That was December 14, 1946. We were run­ning around that morn­ing and some­body took our pic­ture. I was almost 22.

Antonelli: [To Jean Bartik:] Hey, where’s that wed­ding pic­ture you had with John Mauchly.

[To audi­ence:] She has a real­ly won­der­ful pic­ture tak­en on her wed­ding day.

Bartik: Those are the six pro­gram­mers, tak­en in 1986. Kay had a par­ty for us in down­town Philadelphia. To the left is Ruth Lichterman and then John Holberton, who was our man­ag­er. Then Fran Bilas in the black dress, and behind her is Betty Holberton. And then Marlyn Wescoff and behind her is Homer Spence. And then me and then Kathleen. 

Maureen: Just for inter­est, they were up in the library ear­li­er, and they were iden­ti­fy­ing some pic­tures that we have in stores here. They had names of peo­ple that we’d nev­er had on file; it just said Unknown Unknown, and they’ve actu­al­ly been fill­ing in some of our archive infor­ma­tion. Oh, here’s that wed­ding picture.

Antonelli: Yes, this is the wed­ding picture.

Bartik: This is John Mauchly [record­ing skips;]

Antonelli: …and who’s that all the way on the left?

Bartik: Well, he’s Al Kozak, but he was­n’t involved

Antonelli: Maybe he was a friend of Bill’s.

Bartik: Oh, the oth­er one was my hus­band. [laugh­ter]

Maureen: I’ll just stay away from that one.

[Part 6 begins here, appar­ent­ly after some unknown gap]

Bartik: …you know, all the math rou­tines and stuff. So we were busi­ly doing those things. And so John said he felt that we need­ed an eas­i­er way to pro­gram rather than use machine lan­guage, although the UNIVAC had a lot of mnemon­ics, so it was­n’t ter­ri­bly hard to pro­gram in machine lan­guage. So he had the idea—and she was brought in to work on the assem­bler lan­guage. I nev­er worked direct­ly with Grace [Hopper], actu­al­ly. I knew she was there.

Maureen: [To Kay Antonelli:] Did you get—?

Antonelli: I just knew her social­ly, when she would come vis­it the house and stuff. But I nev­er worked again in the com­put­er field.

Maureen: After that. Betty Holberton.

Bartik: Well, Betty Holberton, I would have to say… Now, I don’t know how you work, but the way we worked at that time were in teams. And Betty and I worked as a team to devel­op the tra­jec­to­ry, and she was the best…well, she was my best friend that I’ve ever had in my life. And she also was the best part­ner I’ve ever had in my life, because we worked togeth­er on the tra­jec­to­ry. And the thing about her that always fas­ci­nat­ed me was we would be work­ing on a prob­lem and could­n’t fig­ure out how to do some­thing, so we’d work and we’d go home. The next morn­ing she’s day, You know, I was think­ing this morn­ing how to do this.” Now, she works in her sleep. It was absolute­ly amaz­ing to me because her log­i­cal mind…because it just hap­pened too often for that to be an acci­dent. She had to be doing log­ic in her sleep. So yeah, she was a fab­u­lous pro­gram­mer. And she did the con­sole for UNIVAC I. It was huge. So basi­cal­ly she was like an ergonom­ics engi­neer for the con­sole. A lot of peo­ple worked on it, but she did more work than any­body on the UNIVAC code, the instruc­tion set, test­ing the instruc­tion set.

Maureen: Lila Todd Butler. Did you get to work with her?

Bartik: I worked for her.

Maureen: She was your man­ag­er, right?

Bartik: She was my man­ag­er when I first came. She was head of the group where we did the cal­cu­la­tions and the tra­jec­to­ries. So I worked for her a couple—

Antonelli: She was a pro­gram­ming supervisor…I mean, for cal­cu­lat­ing the trajectories.

Maureen: Was she also a com­put­er and then a programmer?

Bartik: She became one lat­er. She was­n’t one of the ENIAC group. But she worked on the ENIAC after it went to Aberdeen.

Maureen: So she became your super­vi­sor even though she did­n’t have—

Antonelli: She nev­er was our supervisor.

Bartik: No, when I first came there, I was hired to do these tra­jec­to­ries. So she was my super­vi­sor when I did the tra­jec­to­ries, but then later…I mean that was only for two months. Very bright gal.

Maureen: Okay. Somebody out of his­to­ry, Ada Augusta Lovelace. We’re won­der­ing if you knew about her back then. [Bartik and Antonelli laugh] I guess I hit a nerve.

Antonelli: That takes us back a time.

Maureen: Yeah, I know.

Bartik: I don’t remem­ber her very well, do you?

Maureen: I hope not. We were won­der­ing if you had known about her and known about the work she had done.

Bartik: Oh, yes.

Antonelli: No, no. We did­n’t know any­thing at all about—

Maureen: Not at the time.

Antonelli: Not at the time. As a mat­ter of fact, the fel­low that she worked for who tried to build that com­put­er over in England. What’s the name?

Bartik: Babbage.

Maureen: Babbage.

Antonelli: Yeah, Charles Babbage. [crosstalk]

Maureen: Another one of those non-romances.

Antonelli: Eckert and Mauchly did­n’t even know about Charles Babbage. I mean, nobody was talk­ing about com­put­ers. It was­n’t writ­ten up in the ency­clo­pe­dias or any­thing like that. He was like a flash in the pan way back in 1870, or ear­li­er than that. And so Ada Lovelace only devel­oped lat­er, I think, when they were look­ing for a name for a com­put­er lan­guage or something.

Maureen: She actu­al­ly developed—

Antonelli: But very lit­tle was known about her.

Maureen: Who were your role mod­els at the time? Or were you just cre­at­ing it on your own?

Bartik: Well, my role mod­el as a woman, or as a person?

Maureen: As an engineer.

Bartik: As an engi­neer? No, I did­n’t have a role mod­el as an engi­neer, but I mean I had a role mod­el. I had an aunt who got out of Missouri. [laugh­ter]

Maureen: It worked. That was a good one.

Bartik: But uh, well John and Pres. Please. You can’t imag­ine what it was like. I worked for John and I worked for Pres

Maureen: I’m not going to let her talk about John because she mar­ried him. Why don’t you talk about John?

Bartik: Well, John Mauchly, I’ll tell you. I adored the man. 

Antonelli: She could prob­a­bly tell you more about him, because she worked for him.

Bartik: He was the most charm­ing per­son I think I’ve ever met. You would go up to John Mauchly, if you approached him and you quot­ed a lime from Alice in Wonderland, he’d fin­ish it. I mean, you could talk to him about any­thing in the world, and he was so respon­sive to you. You know, a lot of peo­ple you say hel­lo and they go, Uhhhh.” Not John Mauchly. You go up to John Mauchly and you say hel­lo, I mean, he’s made your day. And cre­ative, and he always pushed us, because I real­ly learned to pro­gram the ENIAC from John Mauchly. The first time I met him, when we came back we were sit­ting there with these block dia­grams, and we were on the sec­ond floor of the Moore School and they were putting on a third floor. So they were ham­mer­ing and oh, it was hideous. So Betty and I were sit­ting there try­ing to fig­ure out the accu­mu­la­tor block dia­gram, and this man walked in the room, and he walked all around the room, and he did­n’t say a word to us. And then he came over and he said, You know I’m just check­ing to see if the ceil­ing’s falling in.” 

Well, that’s the first time I’ve ever even seen John Mauchly. So then he intro­duced him­self. Well, we were thrilled. We said, Oh my god, we’re hap­py to see you. Tell us how this accu­mu­la­tor works.” So then his office was next door, so we used to go in every after­noon and get our ques­tions cleared up.

He was a born teacher, and a talk­er. Both Pres and John were teach­ers and talk­ers. You know, if you’d ask them about some­thing, they did­n’t just give you a yes or no answer. They told you all about it, so you were always in a learn­ing situation. 

And Pres…

Antonelli: I’ll tell about Pres.

Bartik: Oh, you talk about Pres. But I adore Pres, too.

Bartik: Want to jump in about Pres Eckert?

Antonelli: When I first came to Moore School in 1942, Pres Eckert was a grad­u­ate stu­dent still, at the time, work­ing for his Master’s degree, and it was because he had been teach­ing John Mauchly in a par­tic­u­lar class that they had actu­al­ly met each oth­er and then John was hired to teach at the Moore School.

But Pres at the time that I came to the Moore School, and short­ly after I came to the Moore School, about two weeks lat­er, I was assigned to work on this dif­fer­en­tial ana­lyz­er, which was in the base­ment of the Moore School. But it was the only air con­di­tioned room in the whole build­ing. And so every­body that had any pri­or­i­ties what­so­ev­er would come down to the ana­lyz­er room to cool off. That was a hot sum­mer that year. And at that time, Pres Eckert and a group of oth­er engi­neers were work­ing doing radar work on the roof of the Moore School. And they would be up there…they’d actu­al­ly strip down till they had noth­ing on but shorts, work­ing up in the hot sun that summer. 

Pres Eckert would come down with them when they came down to cool off in the ana­lyz­er room, and he would still have on his white linen shirt, mono­grammed, and a black neck­tie, and those long black pants that he ever wore. He nev­er dressed down. I don’t think he ever had any infor­mal clothes. And one time, he and I were writ­ing a paper togeth­er. This was years and years lat­er, and John Mauchly had already died. He died in 1980. And Pres and I were writ­ing a paper togeth­er. So at the time, I start­ed telling about my first rec­ol­lec­tion of him, about the white shirt and the black neck­tie and so on. And I said, Didn’t you ever have any oth­er col­ored shirts? Any col­ored shirts?” 

And he said, Yes, I had some col­ored shirts.” 

So I said, How come you always wore white shirts?”

And he said, Well, that’s what my moth­er laid out for me every morning.”

So the next—

Maureen: Oh, the analo­gies are just racing.

Antonelli: I mean, he just did­n’t live in the world—

[Video cuts out]

Further Reference

WITI Hall of Fame pro­file on the ENIAC Programmers.

Prior to this even, WITI auda­cious­ly asked [Jean and Timothy Bartik], as a writ­ing exer­cise, to fill in the Hall of Fame nom­i­na­tion form once again.” [PDF]

Oral his­to­ries with Jean Bartik at the Engineering and Technology History Wiki, and at the Computer History Museum (2008).

The three WSJ arti­cles by Tom Petzinger: