Historical Document · 1983
The Computer Museum Report vol. 4, Spring 1983 — incl. “D. H. Lehmer's Number Sieves”
Includes R. Rubinstein's article on D. H. Lehmer's number sieves and their reproduction for The Computer Museum.
Historical Document · 1983
Includes R. Rubinstein's article on D. H. Lehmer's number sieves and their reproduction for The Computer Museum.
| THE COMPUTER MUSEUM REPORT SPRING/1983 “In competition with the IBM 7094 [it] makes a good showing, especially on economic grounds. It is roughly 10 times faster than the 7094, and costs about 2 cents an hour to operate.” D.H. Lehmer 1 The Director’s Letter Gwen Bell 2 D.H.Lehmer’s Number Sieves Richard Rubinstein 5 Inside “The Soul of a New Machine” Tracy Kidder and Tom West 8 Recollections of the Watson Scientific Laboratory, 1945-1950 Herbert Grosch 13 Museum News Field Trip to North Bay Pray, Mr. Babbage... Founders Program 16 Computer Museum Slide Collection 17 Spring “Bits and Bites” PHOTO CREDITS: Robert Bonifield, Cover, pp. 2, 3, 4; David Bromfield pp. 3, 12, 14, 15; Data General Corporation p. 5; IBM archives pp 8, 9, 10, 11, 12; MITRE Corporation p. 13. THE COMPUTER MUSEUM The Computer Museum is a non-profit, public, charitable foundation dedicated to preserving and exhibiting an industry-wide, broad-based collection of the history of in- formation processing. Computer history is interpreted through exhibits, publications, videotapes, lectures, educational programs, and other programs. The Museum archives both artifacts and documentation and makes the materials available for scholarly use. The Computer Museum is open to the public Sunday through Friday from 1:00 to 6:00 pm. There is no charge for admission. The Museum's lecture hall and reception facilities are available for rent on a prearranged basis. For information call 617-467-4443. Museum membership is available to individuals and non-profit organizations for $25 annually and to businesses for $125 annually. Members receive the quarterly Report, invitations to all lectures and spe- cial programs, new posters, and a ten percent discount in the Museum store. A Founders program is in effect during the initial two-year period of the Museum, until June 10, 1984. During this period individuals and non-profit organizations may become Founders for $250 and businesses and chari- table Foundations may become Founders for $2500. Founders receive all benefits of mem- bership and recognition for their important role in establishing the Museum. THE COMPUTER MUSEUM REPORT (ISSN 0736-5438) The Computer Museum Report is published quarterly by The Computer Museum, One Iron Way, Marlboro, MA 01752. Annual sub- scription is part of the membership of the Museum ($25 per year for individuals and nonprofit organizations and $125 for corporations). The purpose is to report on the programs and exhibitions of the Museum. The con- tents of The Computer Museum Report may not be reproduced without written consent. The Museum Staff is responsible for the con- tents of the Report. The opinions expressed do not necessarily represent those of The Computer Museum or its Board of Directors. The design and production of the Report is done by Benson and Clemons. BOARD OF DIRECTORS Kenneth H. Olsen, Chairman Digital Equipment Corporation Charles W. Bachman Cullinet Software Systems C. Gordon Bell Digital Equipment Corporation Gwen Bell The Computer Museum Harvey D. Cragon Texas Instruments Robert Everett The Mitre Corporation C. Lester Hogan Fairchild Camera and Instrument Corporation Theodore G. Johnson Digital Equipment Corporation Andrew C. Knowles, III Digital Equipment Corporation John Lacey Control Data Corporation Pat McGovern ComputerWorld George Michael Lawrence Livermore Laboratories Robert N. Noyce Intel Brian Randell University of Newcastle-upon-Tyne Edward A. Schwartz Digital Equipment Corporation Michael Spock The Children's Museum of Boston Erwin Tomash Dataproducts The Honorable Paul E. Tsongas U.S. Senator from Massachusetts STAFF Director Gwen Bell Exhibits and Archives Jamie Parker, Coordinator Susan Kaelin John McKenzie Bill Meany Beth Parkhurst Meredith Stelling Programs Christine Rudomin, Coordinator Administration Geri Rogers, Coordinator David Bromfield Debbie Sterling Store Carole Strecker, Coordinator Sally Askins Linda Davidson Merle Insigna The Computer Museum One Iron Way Marlboro, Massachusetts 01752 617-467-4036 © 1983/THE COMPUTER MUSEUM THE DIRECTOR’S LETTER The Computer Museum Report now has an ISSN (an International Stan- dard Serial Number), another benchmark in our development. The purpose of the publication is to report on the Museum's activities with the material in each issue derived directly from our own projects, thus providing a record and a means of communication to our worldwide membership. This issue only covers a part of our fall program of three lectures, seven Sunday afternoon talks, the production of “Pray, Mr. Babbage . . .", the members’ field trip to Canada, a variety of additions to the collection, and a number of new prod- ucts available in the Museum store. Both the Museum and the Report have seasonal rhythms. Writing this letter during our January-February hibernation, I’m recalling all the wonderful activities of the fall season and anticipating Gene Amdahl's talk that starts our spring season on March 10th. Derrick Lehmer opened the fall lecture series October 7th, 1982 and dedicated the exhibit of his sieve machines to the Museum. No sooner was the exhibit mounted than The Scientific American sent a photographer to shoot these unique machines for their November issue. As a result the Museum was given the excellent photos that appear on the cover and in the first article of this Report. On October 8, 1982, the Museum’s first excursion took off for North Bay, Canada to see the AN/FSQ-7. By the time you are reading Gordon Bell's trip report (page 13), the decommission of the Q-7 will have commenced and we will be puzzling over how to fit a representative exhibit of it in the Museum. The Fall also inaugurated Bits and Bites, a series of informal Sunday afternoon talks, ranging from technical presentations to reminiscences about the everyday use and development of computers. These complement the formal lecture series focusing on significant events in the evolution of infor- mation processing. The session “Inside the Soul of a New Machine” with Tom West and Tracy Kidder (reported on page 5) provided an almost anthro- pological view of the species “computer engineer.” Tracy was on the spot to record how Tom behaved at breakfast, as well as at work. In the Spring, the reminiscences of the group that designed the first computer space game will tell the Computer Museum audience why they ever tackled such a project and also provide insights into the everyday environment of a university computer center in 1962. Often the lectures and talks lead to changing or further developing an exhibit. Herbert Grosch's lecture on IBM’s SSEC at Columbia's Watson Laboratory (extracted on pages 8 to12), the two articles on the SSEC in the October 1982 Annals of the History of Computing, and a flurry of letters to the Museum convinced us to add the SSEC to The Pioneer Computer Timeline. With a start up date in January 1948, the SSEC did comparable work to the ENIAC and preceded the EDSAC and Manchester machines. Feedback from members and visitors also helps the Museum preserve the accuracy of its exhibits. Our exhibit coordinator Jamie Parker keeps an ongoing list of text corrections to make. The original text on the Pioneer Computer Timeline was put up crediting “George” Aiken for the Harvard Mark I, the corrected version with an error from the typesetter read “Harvard” Aiken and finally we got Howard's name correct. Several letters pointed out the inaccurate placement of AmdahI's constant on the cover chart of the Fall Report. The corrected version is reprinted on this page and is also available in color in the slide sets listed on the back page. For longer than I am pleased to report the specifications on the Whirlwind read that it used 150,000 kw of power. A number of visitors took great pleasure in drawing this error to our attention, and in some fanciful moments, I've thought of the deliberate intro- duction of errors providing ongoing “treasure hunts” for mistakes. But this added confusion is hardly needed. Our goal is the presentation of the clearest picture possible of the history of information processing to provide a perspec- tive of the present and future as if viewed through a rear view mirror. Gwen Bell * Director Corrected version, cover Winter 1983 Report. The Computer Museum Report/Spring 1983 1 2 The Computer Museum Report/Spring 1983 j ] : | 3 This Photoelectric Number Sieve, built by Dr. Lehmer in 1932, performs 300,000 tests per minute using 30 gears arranged tangentially. Each gear has a number of holes equal to a multiple of a prime. For any problem, all holes that do not represent solutions are plugged with toothpicks. A solution is found when light, originally supplied by an automobile headlight lamp, passes through all 30 wheels. A photo- electric cell detects this brief flash of light, and a vacuum-tube amplifier multiplies the resulting signal 700,000,000 times to stop the motor. = D.H. Lehmer’s Number Sieves Richard Rubinstein Richard Rubinstein, a human factors engineer at Digital Equipment Corpo- ration and an active volunteer at The Computer Museum, compiled this arti- cle, based on his experience planning and supervising the Lehmer Number Sieves exhibit. Dick Rubinstein (left) and Derrick Henry Lehmer (right) pause in front of Lehmer’s Photoelectric Number Sieve after Lehmer’s dedication of the ex- hibit, October 7, 1982. Derrick Henry Lehmer is the son of a leading American number theorist, Derrick Norman Lehmer (1867-1938). D.H. Lehmer designed and built the number sieves shown in the accom- panying photographs of the Museum's Number Sieve exhibit to further his own research in number theory. He pre- sented the number sieves to the Com- puter Museum, and on October 7, 1982, he dedicated the exhibit with a lecture on the sieves. Videotapes of his lecture are in the Museum archives. As a user of ENIAC, Dr. Lehmer was in charge of operations, maintenance, and much of the trouble shooting. Dr. Lehmer continues to use computers, such as the Cray 1 and ILLIAC IV, to support his work with number theory. He is professor of mathematics Emeri- tus at the University of California, Berkeley. The number sieve idea dates back to the Greek Eratosthenes who lived in Alexandria about 230 B.C. His sieve provided a way to find prime numbers by removing composite numbers from a list of the positive integers. The first mechanical realization of a sieve process was over 200 years ago, by Karl Hindenberg, a German, and Anton Felkel, an Austrian. They in- vented what is now called the stencil method. Using this technique, Felkel produced a compilation of the factors of the integers up to 408,000, published by the Austrian government in 1776. Few copies were sold, and those remaining were scrapped and used for making cartridges for the war against Turkey. These sieves were specialized, useful only for finding primes and for finding the factors of composite inte- gers. A more general approach, the strip method, used by A.M. Legendre in 1794, allowed searching for numbers with far broader properties, but was still a slow and error-prone manual technique. The strip method was the best available technique when D.H. Lehmer developed his first machine. Dr. Lehmer built the first electro- mechanical number sieve in 1926. It used 19 bicycle chains, and performed a process similar to the strip method. The rotating chains simulate the effect of the shifting strips of paper, and have the major advantage that no upper limit of the search need be established at the outset. The machine can be left running indefinitely in search of numbers with the desired properties. In 1932, Lehmer built a far faster machine using 30 tangential gears, capable of perform- ing 300,000 tests per minute. Later, he also built a sieve that used 16mm film, and another with vacuum tubes and delay lines. These machines did effi- ciently and inexpensively what no com- mercial calculating equipment then available could do at all. This small sieve, employing 16mm film as the computing elements, was built as a complement to the large photo- electric machine. It was used for rela- tively short problems, and performed about 3000 tests per minute. Easier to set up than the photoelectric model, it could be supplied with as many as 18 film loops. The Computer Museum Report/Spring 1983 3 TE These stencils were used by D.N. Lehmer (D.H. Lehmer's father) to pro- duce a table of factors up to 10 million. A great deal of kitchen-table work was performed by the whole Lehmer family in order to create these card- board pieces. Number theory seems to be a family activity with the Lehmers. Dr. Lehmer is now working with his grandson on a new machine using integrated circuit technology. 4 The Computer Museum Report/Spring 1983 The Bicycle Chain Sieve, built in 1926 but later destroyed, is being replicated at the Computer Museum by Roberto Canepa, Andrew Kristoffy and Richard Rubinstein. This photograph shows the Museum's working model before the installation of micro-switches and a counter. The original 19-chain ma- chine could perform 3000 tests per minute. It was used to factor the number: 9999000099990001 which is a factor of 10° + 1. The ma- chine ran for about two hours, finding the solution: 1676321 * 5964848081 } i |: Constructed in 1966 as an unsponsored educational project of the Departments of Mathematics and Electrical Engi- neering at the University of Califor- nia, Berkeley, the Delay Line Number Sieve performed | million tests per second. Navy surplus delay lines store the acceptable remainders for each modulus and circulate them synchro- nously. The machine has an idle mode in which all of the delay lines are con- nected in series and the data recircu- lated, so that the machine may be “stopped” and loaded. “In competition with the IBM 7094 [it] makes a good showing, especially on economic grounds. It is roughly 10 times faster than the 7094, and costs about 2 cents an hour to operate.” D.H. Lehmer Inside “The Soul of a New Machine” Tracy Kidder and Tom West Tom West: I think it's probably important to point out that Tracy and I have never done this before . . . and will probably never do it again! I am still doing com- puters and Tracy is still writing books. Neither of us has yet become PR. junk- ies, but that may change after looking at such a large group of people who seem to think we have something to say. Tracy Kidder: We would like to find out what you are interested in and try to answer your questions. Q: How did you two first get together? Did you start writing at the time the project was started? Kidder: About six years before I started, I had been a journalist writing for the Atlantic Monthly. I went in to my editor one day because I had nothing to do and asked him what I should do next. He said "Look into computers.” I asked where I should start and he told me to go see West, whom he knew. So I did and then one thing led to another. West: It was a little more complex than that. Tracy decided that he was going to write “The Book” about all computers. He went to the library and found that there had already been many stories written about the whole world of com- puters. Then it was going to be the whole world of mini-computers, which I suspect even today is too big a story to tell. It kept narrowing and narrowing, until, rather accidentally it became focused on a single machine. Q: What was it like for you, as someone who is used to working with machines in a crisis situation, to be shadowed by a person with a notebook? West: There are some real advantages to having a writer on a project. Tracy was about the best early warning sys- tem you could imagine. I would suggest that every program manager hire some guy from the street, teach him how to write on steno pads and have him walk around listening to people. The young Tracy Kidder (left) and Tom West (right) discuss The Soul of the New Machine during the Museum’ Bits and Bites series. engineers walk around thinking that this must be an incredibly important program because somebody is writing about it. One disadvantage is that you could wind up with the most well- recorded failure in history. Just Tracy looking at the thing actu- ally changes the way it works. There is all this uncertainty involved. It's not the way it would work under normal circum- stances, it has been perturbed. It is a little disarming to wake up in the morn- ing, go down to breakfast and find Tracy Kidder interviewing your wife and two daughters with a steno pad. Kidder: Most good journalists, and I certainly aspire to be one, think of them- selves as good anthropologists who come in to observe the customs of the natives and don't want to change those customs in the process. Just the fact that you are there changes things. But, if you are never there at breakfast then you would never know what happens at breakfast. Q: What sort of reaction did you get from Data General and from people in the book, after the book came out? West: For me, it was like three years of psychoanalysis embedded in about twenty-four hours. I hadn't read the book before Tracy went to the printers with it and I don't think many of the other people had either. It was a shock then, to all of a sudden see a whole piece of your life in print, a piece that maybe you would have chosen to rewrite very care- fully. I don't think Data General was cer- tain, even after reading the book, what the public reaction was going to be. It was a fairly accurate representation of what they do every day for a living. And I don't think that the story is unique. Otherwise, why do so many people read the book? Why do so many people buy the book and send it home to Mom to explain what they do? And why do they give it to their girl friends to explain that this is the reason they are not home until eleven o'clock at night? The general answer is that the everyday life of designing computers had been pretty opaque. But from my point of view the re- viewers tell you what you are supposed to think about the character. The Computer Museum Report/Spring 1983 5 SS Kidder: That's right, you have been called Captain Ahab, The Prince of Darkness, Machiavelli, Tom Swift. . . West: Gatsby and Horatio Alger. It's been kind of eerie. Q: The paperback book says it is soon to be made into a major motion picture. What is exciting to the lay person about high technology? Kidder: think this is kind of fun, frankly. Columbia wanted to tum this into a movie, one way or another. As to why they find it interesting, I don't know. I found the story interesting and I rate as a lay person. At one time during my research, Data General had some un- derstandable fears about trade secrets. West pointed out in a rather sardonic way that I would never learn enough to be able to convey a trade secret. West: When people ask why D.G. al- lowed a Pulitzer Prize winning book to be written about their machine, that is not what they agreed to. A guy was going to write a little piece about a couple of guys sitting down in a lab sort of dreaming up their own thing. Even at printing time there was really no way of knowing that there was quite so much interest in Tracy's book. Q: What are some of the advantages of a closed management style versus a more open one? West: In support of this sort of closed management style, I knew more about what was going on in every single one of those organizations than any manager at that level in most companies does. It's not a question of not having the informa- tion, it's a question of not having the style. If somebody walks into your office after five minutes of staring at a sheet of paper and asks: “Should we make this register sixteen bits wide or thirty-two bits wide?” you know the answer to this question if you have been through it over and over and over again. So you can tell him the answer, and he goes back and puts it in. Then he sits around for another five minutes before coming back with the next question. In some sense that is sort of what happens. If you can send him back to his office and he thinks a little while longer, ten minutes, fifteen minutes, or a half an hour, and then he gets a little scared because he’s got to have the right an- swer by the time he goes back in again, you extend his span of attention before he is finally ready to go and scream for help. It seems to me that that is one of the most valuable lessons that you can 6 The Computer Museum Report/Spring 1983 teach a kid who is coming into this business right out of school. A span of attention of five minutes or even a half an hour is not really going to do it. There are some problems that only yield after three hours of just staring at them. I'm sure all of those who have been doing design work have seen exactly that same thing happen. It is the ability not to give up even after two hours of trying to stare something down. Part of that is what is embedded in that whole style of trying to make the easy answers dif- ficult to get. Q: It is very clear at the end of the book that you couldn't keep a good team together, but it is also sort of sad that so many of them end up leaving Data Gen- eral, West: At the end of this program, an awful lot of people, not only left the program, but also left Data General. This is not a totally anomalous phenom- enon to D.G. People don't leave projects in the middle, they leave them at the end, with an incredible postpartum de- pression at the end of something that they poured so much adrenalin into. They go home and with more time on their hands. They begin to ask them- selves why the heck they ever did it in the first place and would they ever be willing to do it again. If the answer is yes, then what are the odds of their doing it successfully again? Is it a coin toss, fifty-fifty each time, or do the odds keep piling up against you? I read in Time Magazine that sta- tistically I'm burned out and washed up at forty-three. I think there is going to be not only a shortage of engineers in the next ten years but a shortage of engi- neering managers with enough street sense to be able to manage all the college graduates who are graduating with all kinds of notions about things that can be done but quite possibly shouldn't be done. Just like all these people who are putting 68000s on a board and calling it a computer. There is going to be a great need for people to be able to see the way technology is going and be able to manage these people coming out of school. In this machine we had a straight- forward design. We were highly lever- aged on PALs because it provided quite a bit of logic compression. Almost every- one drew out these PALs, except the guy who was designing the system cache. He was doing it all on four by five file cards. Every week or so I'd ask him how he was coming on his design. He'd show me this big deck of file cards with Bool- ean equations written on them and a whole room full of Karnaugh maps and things to try to reduce these down to something that made sense. There was no way of telling where the guy was. Design review time was coming and all the managers were getting an- xious. The big deck of cards was still there, and he was still worrying about the various mapping things that he'd learned in college. We really came peril- ously close to firing that guy. But finally when the design review day came, it was a deck of cards with a well-defined set of Boolean expressions for each one of those PALs. We used one hundred and sixty of them in the machine and about forty of them on the system cache. He built the thing, and it worked the first time. At that stage of the game I said to myself; thank God I didn’t fire him. Q: Would you attribute the intimidation some of your subordinates felt toward you to their inexperience? West: In the first place, I would sure hate to defend what I did four years ago. At some level in the organization you ex- pect to have people who are going to fight back, regardless of what the issue is, if they feel strongly about it. If they don't feel strongly about it, then they had better go away and do it your way. Q: The book mentions the fun part, the engineers doing the design, but I don't remember hearing about the paper work and bureaucracy, meetings etc. Did you find some way around that? Kidder: You know the old joke about the bum who is looking for his quarter under the street light, even though he dropped it half a mile away because the light is better. This book doesn't take on, in any detail, the software part of the project which is at least half. There were some things that I saw more of than others. I was also able to drag a certain amount of that information out of Tom. Then one day I said to Tom: “I don’t know much about what you've been up to lately,” and Tom said, "Oh you noticed.” I always thought that the salient characteristic of this team, which I thought was charming, was that no one ever seemed to take pert charts at all seriously. A nucleus had built comput- ers before and they knew what was required. They knew, I think, that what was required could simply not be em- bodied in a pert chart. Q: If you could do this over again, would you fight harder for a mode bit? West: No, quite the contrary. In hind- sight I thought that was a really spec- tacular decision. Tracy said half the work may have been software where I would guess eighty percent of the work was software. Without the mode bit we could drop sixteen bit programs on that machine and run them. For a long pe- riod of time that was exactly what we were doing. We also had three or four hundred thousand lines of diagnostic code, all of which would have had to be rewritten. So for the time being, having that absolute compatability was the only thing that could get it there on time. The point is that the machine doesn't have a golden moment of when you all of a sudden stop implementing the mode bit and then move on to a different identity. Mode bits tend to in- crease in a geometric fashion, and I've seen a lot of machine families that are now going into their two-to-the-sixth mode bit. Intel is going to solve all their incompatability problems at the chip level with mode bits and it is going to go to huge numbers, I think. Q: I wonder if you could say something about Ed DeCastro's involvement. The novel makes it seem like he was very rarely checking on the project. West: That is one of the most diffi- cult questions to answer because once again I think most of the people who were working on the project itself would as- sume that I had no visibility, barricading myself in my office. Mr. DeCastro is more than likely to let something go with benign neglect, assuming that local management would find a way to solve its problems. He was certainly not in- volved on a daily or weekly or even monthly basis. Q: What about Carl Carmen, who was the vice-president of engineering at that time? West: He was involved during the pro- gram, in making sure that the envi- ronmental issues were taken care of, that the PC shop worked at the right speed and that things didn't get lost in the mill. He also barricaded the team against the rest of the organization. Q: Since you weren't able to tell us much about software, those of us who are in software felt a little cheated. Could you tell us what the organization of the soft- ware development's side was? West: When Tracy talked about looking under the street light for the quarter it is because hardware tends to be a lot easier to see, it has a lot more of a visual effect than writing code does. The reason for having the sixteen bit identity in the machine initially was because we didn’t believe that we were going to have any software at all. There was a guy who decided that he could take our existing sixteen bit operating system and by taking it module by mod- ule convert it and run it initially in just a couple of the rings with only a couple of the features, and then incrementally get there over a period of a year and a half. He put together a team of twenty volun- teers all signing up for a kamikaze mis- sion because nobody really believed that it could be done. Then he decided that we were going to have the software to announce at the same time that we had the hardware to announce. All we had planned to announce was thirty- two bit hardware, which would get some pressure off our back and point to futures. He managed to doit. The reason that he managed to do it was that he didn't decide okay, I've got a clean sheet of paper and I'm going to develop an O.S. all the way from the ground up, based on the first principles of computer science. The relationship was quite close. In the final analysis, the only thing we really had to de-bug in the thirty-two bit part of the machine was the operating system which had already been run through thirty-two bit simula- tors. At that stage of the game we were using system software to de-bug the hardware. Q: What happened to the competing design? West: The competing design is still alive. It's always difficult to know the answer to “What if?”. It would be naive to suppose that D.G. was only working on extensions to the Eagle family of machines. It would also be naive to assume that we are just pragmatically going to follow technology and wander along incrementing the existing product line. The question is when, not whether. Q: The book mentions a saying in Mr. West's office that anything worth doing is not necessarily worth doing well or something along those lines. Kidder: One of the things that I began to learn about engineers is that they are aesthetes as much as they pretend to be something else. I've seen this most vividly in the intersection of engineers and non-computer scientists. The engi- neer talks about technical symmetry and the scientist says I just want some- thing that works. I think that that was what that piece of cryptic puzzling ad- vice was. How did it go? Not everything worth doing is worth doing well. West: I am very comfortable with that notion. I suspect that there are more people who fail in our industry because they try to do it perfectly, as opposed to doing it on time and on cost. Q: What are you doing now, respec- tively, and secondly, are you (Tracy) sick and tired of computers? West: We are still basically doing the same thing—designing machines—at Data General. The book portrays people leaving, which was true at the time Tracy had to go to press, but since that time a large number of people did stay. All those people have formed a nucleus to build multiple different machines, going in different directions, they build bigger ones, smaller ones, faster and cheaper ones. Kidder: I'm digging out from under. I'm writing some articles about atmospher- ic research. To be honest, I'm a little tired of my book. I put it on my shelf and won't read it again for years. I think I know what's wrong with it. In some sense, writing a book is like building a computer. There are rewards but one of the main ones is that Sisyphean one that if you do one you get to do another. So, I have an opportunity now to write a better one. Extracted from a talk by Tracy Kidder, author of The Soul of a New Machine and Tom West, the chief designer of the machine, Data General's Eagle, in the Museum's “Bits and Bites" series, given October 17th, 1982. THE SOUL OF A NEW MACHINE Tracy Kidder autographed copies of The Soul of the New Machine for the Museum. Remaining copies of the autographed copy are available for $18.00 (including shipping) from the Museum store. The Computer Museum Report/Spring 1983 7 as Recollections of the Watson Scientific Laboratory, 1945—1950 Herbert Grosch The heroes of this story are Thomas Watson Sr. and Wallace Eckert. Tom Watson Sr., when I first met him, was in his seventies. In the 20's, 30's and 40's, he supported what he thought of as scien- tific research, but what we call applied research. At that time, this kind of sup- port was uncommon for even the most advanced American industries. Wallace Eckert's prime life interest was the theory of the motion of the moon. Astronomy provided one of the earliest groups of people who had sub- stantial problems which had to be solved with computation regardless of the form, whether it be pencil and paper, logarithm tables, or with the most ad- vanced modern solid-state computing devices. In the thirties, a major problem was that crossing the Pacific required the measurement of both longitude and latitude. The chronometers were in- accurate and a very accurate clock In 1945, IBM remodeled this fraternity house and donated it to Columbia University for the Watson Laboratory. The first concept of the SSEC was formulated in this building and the group that eventually built the NORC (Naval Ordnance Research Calculator) worked on the upper floors and in the basement. 8 The Computer Museum Report/Spring 1983 was needed, Many many years ago the motion of the moon among the stars was suggested as a clock. As a conse- quence, a great effort was put on the determination of the proper position of the moon. One of Babbage’s interests was in building astronomical tables helpful for navigation. Even today, peo- ple who are doing algebra on comput- ers, often use the enormously complex and lengthy formulas of the litteral lu- nar theory as an example. Eckert discovered a series of arti- cles by L.J. Comrie, a New Zealander who by that time had penetrated the scientific establishment in England suf- ficiently to become the Director of the Nautical Almanac office at the Royal Greenwich Observatory. These articles explained how Comrie had rented Hol- lerith machines from the British Tabulat- ing Machine Company and had done the calculation of the positions of the moon for every hour or every six hours or something for hundreds and hundreds of years on a mass production all- parallel basis. Inspired by this, Eckert, with the help of the American Astronomical Society of which he was a junior mem- ber and Columbia University where Thomas Watson Sr. was a Trustee, ap- proached the IBM Corporation, a less than one-hundred million dollar busi- ness. They donated some special equip- ment which he installed and called it The Thomas J. Watson Astronomical Computing Bureau which flourished in the late years of the thirties. Then the War threatened. The U.S. Naval Observatory had to manufacture a new publication called the Air Al- manac for the use of navigators crossing the Atlantic by air, especially for bomb- er navigation, when the skies were clear and bubble sextant observations and so on were possible. They had run out of old-fashioned people who could do this with paper and pencil and logarithm tables. Wallace Eckert's war work was the real McCoy—carrying out very com- plicated calculations, the sort that f of, \ ~~ a « This famous photograph of Thomas Watson Sr. hung over the fireplace in the Watson Laboratory at Columbia University for which he furnished the money and a large part of the inspiration. Wallace Eckert strained the professional astronomer. Done, however, in parallel so you didn't have the problem of a complex se- quence, but doing a single simple oper- ation to hundreds or thousands of pieces of data at one time and then moving on. With methods he developed for auto- matic proof-reading, he was able to pro- duce the Air Almanac. Since that time literally millions and millions of charac- ters have been printed by the calcula- tions of such equipment, printed by his automatic printing machine, and proof- read by automatic devices without one single error ever having been detected. In 1941, I became an optical de- signer which involved lots and lots of calculations on Monroes and Fridens. Then, in the early spring of 1945, an announcement in Science stated that Thomas Watson Sr. had called Wallace Eckert back from Washington and had asked him to establish a new scientific computing laboratory at Columbia University. Before he left Washington, I wrote Wallace asking if I could come around in the evenings and try out some of my ideas on optical design on his nice new shiny IBM machines. One of the incentives for beginning the laboratory was that the computational facilities at Los Alamos had run out of capacity. Wallace accepted the charge of setting up a shop of IBM machines at Columbia with the first task to supplement the Los Alamos calculations on the Alamogordo burst. While I only hoped for an invita- tion to maybe around one night a week, instead a little man from the Manhattan Project showed up at my optical com- pany and took me away. I said, “You know you can't do that, there is no such thing as a civilian draft.” He essentially said, “Tell that to General Grove.” The next thing I knew I was an IBM em- ployee. In the rush they forgot to pass me through IBM headquarters. As a result, although I received an identifica- tion card and all that, nobody had paused to tell me that you could not have hair on your face and work for IBM. So I was not only a very early scientific punch card operator and supervisor but lalso was the first bearded sport-coated IBM employee. 1946. Boom. ENIAC. For the first time IBM felt threatened by a development that they had not really foreseen or un- derstood. One of the responses to the ENIAC announcement was the mass production (mass being about twenty units) of the 603 calculating punch oper- ating at around six thousand cards an hour while its electro-mechanical com- petitors did six hundred. IBM had pro- duced it out of their own patents, their own Eccles-Jordan flip-flops and so forth, originating primarily with a gen- tleman by the name of Halsey Dickin- son. But, this card calculator was not enough for either Eckert or Thomas Wat- son Sr. who was incensed that someone would produce something that he didn’t know about and hadn't sponsored. A group of people were brought together to write the specifications for a gigantic new machine, the SSEC. With almost no electronic gear available on the market, the arithmetic units were designed around the standard 25L6 ra- dio vacuum tubes. The design of the SSEC went ahead day and night, seven days a week at the IBM engineering laboratory in Endicott. Along with the electronics, a complete panoply of peripheral equipment was designed: high speed card readers, auxiliary tape punches, card punches, fancy con- sole, storage devices, and a major table look up unit in contrast to the setting switches on ENIAC'’s function table panel. The whole thing was to be installed in beautiful quarters in the IBM world headquarters at the comer of Madison Avenue and 57th Street, since torn down. Since no sizable ground floor space was available, they bought out a store called the French Bootery around the corner on 57th Street. They tore out the shoe store shelving and put power supplies with a gigantic air-conditioning unit in the basement. The equivalent of false floors was created by a raised floor with the enormous amount of electronic cabling under it. The machine was put together at Endicott, ran, taken apart, and moved to New York City. On Mr. Watson's final inspection about two or three days before the open- ing ceremony, he was disturbed by the fact that there were large columns marching down the center of the room. He said, “Everything is lovely, you gen- tlemen have done a beautiful job but I think we should remove those columns." Unfortunately, they held the building up. Nevertheless, the four-color bro- chure which had been printed for open- ing day was recalled and a two color sepia print center-fold inserted showing the machine room minus its columns. After the dedication, Thomas Wat- son Sr. said, “It's wonderful how these people out at Endicott and these people in New York have slaved over this ma- The