 |
 |
 |
 |
 |
 |
|
|||
 |
 |
 |
|||||||
| Copyright © 2012. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page R1
Electronic Brains: Stories from the Dawn of the Computer Age
Electronic BRAINS
OCR for page R2
Electronic Brains: Stories from the Dawn of the Computer Age
This page intentionally left blank.
OCR for page R3
Electronic Brains: Stories from the Dawn of the Computer Age
Electronic BRAINS
STORIES FROM THE DAWN OF THE COMPUTER AGE
MIKE HALLY
Joseph Henry Press
Washington, D.C.
OCR for page R4
Electronic Brains: Stories from the Dawn of the Computer Age
Joseph Henry Press
500 Fifth Street, NW Washington, DC 20001
The Joseph Henry Press, an imprint of the National Academy Press, was created with the goal of making books on science, technology, and health more widely available to professionals and the public. Joseph Henry was one of the founders of the National Academy of Sciences and a leader in early American science.
Any opinions, findings, conclusions, or recommendations expressed in this volume are those of the author and do not necessarily reflect the views of the National Academy of Sciences or its affiliated institutions.
Library of Congress Cataloging-in-Publication Data
Hally, Mike.
Electronic brains : stories from the dawn of the computer age / Mike Hally.
p. cm.
Includes bibliographical references and index.
ISBN 0-309-09630-8 (cloth)
1. Computers—History. I. Title.
QA76.17.H35 2005
004—dc22
2005016583
Cover design by Michele de la Menardiere; photo, copyright by Hulton-Deutsch Collection/CORBIS
Copyright 2005 by Mike Halley. All rights reserved. First published in Great Britain by Granta Books 2005 by arrangement with the BBC.
Printed in the United States of America.
OCR for page R5
Electronic Brains: Stories from the Dawn of the Computer Age
CONTENTS
Preface
vii
Prologue
xiii
1
From ABC to ENIAC
1
2
UNIVAC—Savior of the Census
29
3
Saluting the Moose
51
4
When Britain Led the Computing World
75
5
LEO the Lyons Computer
103
6
So Then We Took the Roof Off
135
7
Wizards of Oz
161
8
Water on the Brain
185
9
It’s Not About Being First: The Rise and Rise of IBM
207
Epilogue
233
APPENDIXES
A
Bibliography and Further Information
241
B
Arithmetic
247
C
Technical Bits
253
Index
257
OCR for page R6
Electronic Brains: Stories from the Dawn of the Computer Age
This page intentionally left blank.
OCR for page R7
Electronic Brains: Stories from the Dawn of the Computer Age
PREFACE
This book is about that brief period just after World War II when the first truly modern computers were developed by a range of pioneering teams across four continents. The book grew out of a series of four 15-minute programs called Electronic Brains on BBC Radio 4 that received an unexpectedly warm welcome from listeners who had had no idea that computer history could interest them. It did that by avoiding too much technical detail, instead exploring what it was actually like to be one of the pioneers in the early computer projects. To get those accounts meant finding surviving members of the pioneering teams in various parts of Britain, America, Australia, and the former Soviet Union, and recording their memories at some length, memories that were invariably clear and vivid.
You can’t fit much of that material into an hour’s radio—about as much as a single chapter in this book—so this version of Electronic Brains is much more than just transcripts of the radio programs. There is more about each of the four stories from the original series, the contributors have more space to speak for themselves and there is more historical context. There is a chapter about the early Australian computers, unjustly overlooked in most accounts. And a major question that arises out of those stories—why IBM became so dominant despite a late entry into the field—is explored in its own chapter.
OCR for page R8
Electronic Brains: Stories from the Dawn of the Computer Age
My first experience of computers moving out of the laboratory and into everyday life came in the early 1970s in a large aerospace engineering company. A shiny new hot-drinks machine replaced the old one on the shop floor and it had a numbered keypad instead of large blue plastic buttons. It had a simple computer inside it, so instead of pushing the button marked “coffee white no sugar” we had to press the numbers “1, 4, 3” on the keypad. It baffled us and for several days we kept getting the wrong drinks, though we were all electronic engineers and we knew that microcomputers were going to find their way into all kinds of products. Within a few years many of us had some sort of computer at home, and they were fascinating to us even though there really wasn’t that much you could do with a Sinclair Z80, other than marvel at it.
Recalling that period helps me understand why all the people I interviewed for this book so obviously enjoyed their work, which was intellectually demanding with long hours and scarce resources. As one of the LEO computer engineers, Frank Land, remarked, “It was very exciting because everything you did had never been done before.” The work was the technological equivalent of climbing Everest or walking to the South Pole, and there was a huge sense of camaraderie, of sheer fun, in those early days, which comes across strongly in the memories of these men and women.
Bringing these accounts together allows certain themes to emerge. One of the most striking is the effect of World War II on the development of the early computers. In some cases the onset of war held back progress, but in most it accelerated it. The war didn’t just bring into being major military projects like the British code-breaking Colossus and the American ballistic computer the ENIAC; there were indirect effects as well. The British professor Sir Maurice Wilkes said the biggest lesson he learned from wartime radar work was “how to get things done.” Across the Atlantic the legendary boss of IBM, Thomas J. Watson Sr., saw that the urgency of
OCR for page R9
Electronic Brains: Stories from the Dawn of the Computer Age
wartime production would force a similar increase in the pace of technological change in his country when peace returned. Another important effect of the war in the United States was the large number of young men who got technical training through the “GI Bill,” which entitled ex-servicemen to a free college education. Many of the engineers working on the early American computers could not have done so without the qualifications they acquired through the GI Bill.
What is also noteworthy is the way the first computers emerged more or less simultaneously in various parts of the globe, in a general climate of technological advancement that spread across the developed world. The great Soviet pioneer Sergei Lebedev may have been spurred on by news reports of the American ENIAC in 1946, but he had been experimenting with the principles of digital computing some years earlier and the machine his team created was all their own design. So there is no simple answer to the question, “Who invented the computer?” It was a process of development across several continents that involved thousands of people. Certainly some individuals made outstanding contributions and others were exceptionally far-sighted, but no one has the undisputed title of “Father of the Computer.” John Atanasoff, with a better claim than most as creator of the Atanasoff Berry Computer (ABC), summed it up well when he said, “There is enough credit for everyone in the invention and development of the electronic computer.”
Atanasoff was quite typical in being a professor of both mathematics and physics. While some of the pioneers were particularly renowned as brilliant mathematicians, like Alan Turing, or physicists, like John Mauchly, nearly all had expertise in both disciplines. It is interesting that most of the early projects were led by teams of two. Atanasoff needed the gifted electronics student Clifford Berry to make his design a reality, just as Mauchly needed
OCR for page R10
Electronic Brains: Stories from the Dawn of the Computer Age
Presper Eckert. In Australia the mathematician and physicist Trevor Pearcey found his electronics guru in Maston Beard. The Russian Sergei Lebedev was almost unique in having the required level of expertise in mathematics, physics, and electronics. Women played significant roles in nearly all the early projects, whatever the country and political system. Female mathematicians, whose employment opportunities had generally been more limited than those of their male counterparts, found their skills in demand and often showed particular aptitude for the new art of programming.
Because of the global coverage and the extensive use of direct quotes in this book I’ve adopted local usage where appropriate. So in the American and Australian chapters I write of electronic “tubes,” while in the British accounts the same devices are called “valves.” Similarly, in the account of the first Soviet computer, I’ve preserved the use of English of the Ukrainian speakers rather than rewriting their testimony in standard English, and I’ve used their preferred spelling of Kyiv rather than the older Kiev. Many of the quotes in this book are taken from recent interviews with pioneers looking back half a century or more, and I indicate this by using the present tense. Extracts from contemporaneous accounts are introduced in the past tense.
It only remains to thank all those who have assisted so generously in making this book possible: Gail Lynch at Granta, who heard a quirky little radio series and thought it might be the basis for a book; Mark Whitaker, who presented that series, conducted the American interviews and cast his historian’s eye over the scripts; all the surviving pioneers of these early projects who granted me interviews and conveyed their experiences so vividly; David Caminer and Frank Land, who read the LEO chapter and corrected some glaring errors; Erik Rambusch, Andrew Egendorf, and Bill Wenning, who did the same for the Rand 409; Viktor
OCR for page R11
Electronic Brains: Stories from the Dawn of the Computer Age
Ivaneko, who led me to the pioneers in the Ukraine—I’d have been lost without him; John Deane, who drew my attention to the early Australian computers and generously gave me his contacts; and above all Doron Swade, who read and forensically scrutinized the whole manuscript despite many other competing demands on his time. Naturally any remaining errors are my responsibility.
Mike Hally
author@electronicbrains.info
OCR for page R14
Electronic Brains: Stories from the Dawn of the Computer Age
mathematical tables that had become widely available in the eighteenth century. This wasn’t just an academic problem. Tide tables and star charts in particular were calculated by hand, perhaps with the aid of slide-rules (invented a couple of centuries earlier), but such tables were notoriously unreliable, with mistakes in calculation, transcription, and printing. These were serious matters in an age when international trade depended largely on maritime transport and errors cost ships, goods, and lives. It is said that Babbage, a great collector of such tables, one day cried out in exasperation that he wished they had been calculated by steam power and then set about designing a machine to do the job.
Only mathematicians could solve the equations that yielded functions like logarithms and sines, but a simplified procedure known as the “method of differences” gave results that were accurate enough when carried out properly. This method could be used by teams of people with only basic arithmetic skills and Babbage realized it could also be mechanized to eliminate the human errors. So he called his calculating machine the “Difference Engine” and started work on it in 1821. Results would be embossed directly on to thick sheets that could be used to print the final tables, thus eliminating the other sources of human error in production. Unfortunately the whole design called for the manufacture of some 25,000 parts, most of them at the limit of materials and machining tolerances of the time. After 11 years, the expenditure of a great deal of government money, and a terminal dispute with his engineer, work stopped. Eventually a small part of the Difference Engine was completed; it worked well (and functions impeccably to this day, according to the Science Museum in London, where it is now kept) and can be regarded as the first automatic calculator.
Although it proved a big hit at dinner parties (Babbage was a great socialite), the incomplete engine didn’t produce the accurate
OCR for page R15
Electronic Brains: Stories from the Dawn of the Computer Age
printed tables Babbage wanted. That was achieved by a father-and-son team, Georg and Edvard Scheutz, who read about Babbage’s efforts and then designed and built their own machine in Sweden. Scheutz senior forecasted in 1833 that just one such machine “would suffice for the needs of the whole world,” probably the first example of a prominent figure conspicuously underestimating the potential of computing. However, the Scheutzes, who built their prototype in just six years with only modest machining facilities, had trouble persuading the world that it needed any such machine at all. This was a pity because, although the prototype was rather basic, it worked and it printed the results. They produced a few more machines to a higher standard of accuracy, which also worked but were rather unreliable and didn’t really produce the expected benefits. Both father and son died bankrupt, six years apart.
Babbage meanwhile had started pursuing another project, his “Analytical Engine,” and this really earned him a place in computer history. During his years of work on the Difference Engine, he had conceived of a far more powerful machine that would be universal in the sense that it would be capable of solving a variety of algebraic equations. Just as he had originally called for mathematical tables to be calculated by steam, he looked again to the technologies of the day for inspiration and to the textile industry in particular. He designed machinery to add, subtract, multiply, and divide, and he called this the “mill”; this was analogous to the arithmetic processor in an electronic computer. Rather than duplicate this machinery all over the engine, he intended the mill to be a single central mechanism that would fetch data from another part of the engine he called the “store.” He came up with a method for the engine to carry out various kinds of calculations by following a set of instructions that defined the equations to be solved. This was long before magnetic tape was invented, but there was a suitable technology, developed for the textile industry: punched cards. These
OCR for page R16
Electronic Brains: Stories from the Dawn of the Computer Age
had recently been invented by Joseph-Marie Jacquard, a Frenchman who used holes punched in cards to control the machines that wove intricate patterns in silk (the idea was also adapted for the mechanical piano, the pianola). So Babbage used punched cards to define the instructions—the “program,” as we would call it—for his Analytical Engine. Thus the Analytical Engine would carry out a series of discrete steps, just as a digital computer does today.
Babbage also gave his engine the ability to choose one set of instructions over another depending on the result of an earlier calculation, what we now call “conditional branching.” This was a remarkable feature, not present in many of the early electrical and electronic calculators a century later. Unfortunately he didn’t build any of his Analytical Engines either, although he did realize during his years of work on their designs (he produced a variety of plans and notes) that he could simplify his Difference Engine. So he sat down and designed Difference Engine No. 2, although this too wasn’t built in his lifetime. Babbage was an example of someone who too easily moved on to the next project before completing the first, and in this respect also he was something of a forerunner to some of the twentieth-century pioneers.
For many years it was generally assumed that the main reason Babbage’s “engines” were never completed was simply that they couldn’t be built using the technology of the day. That was until a team led by Doron Swade, then curator of the computer collection at the Science Museum, was prompted by the Australian historian Allan Bromley to test this assumption by building a replica of one of them. They chose Difference Engine No. 2. Work started in the mid-1980s with the aim of having a functioning machine by December 1991, the 200th anniversary of Babbage’s birth. They limited themselves to the materials and machining standards of the nineteenth century, and for time and cost reasons decided not to
OCR for page R17
Electronic Brains: Stories from the Dawn of the Computer Age
build the printing part of it. Thus began a six-year roller-coaster ride that involved fund-raising, corporate reorganization, seven-day weeks, a key contractor going bust, and much more, echoing Babbage’s original trials and entertainingly recounted in Swade’s book The Difference Engine. But the experiment worked and the result can be seen at the Science Museum today; a trip there is recommended for anyone who really wants to understand Babbage’s engines. The printing and stereotyping apparatus was added later and was working by the spring of 2002. The exercise proved that Babbage’s failure wasn’t due to the limitations of nineteenth-century machining but rather to his dispute with his engineer, Joseph Clement, Babbage’s inveterate tinkering with the design, and above all the fact that he just wasn’t a very effective project manager.
Ever since Babbage’s time it has been assumed that he was driven solely by his search for error-free calculation, but recent research by Doron Swade suggests this is a myth. Rather, says Swade, Babbage saw the engines as a new technology of mathematics able to systematically solve complex equations and compute functions for which there was no analytical law. Babbage even predicted a new branch of mathematics, which came into being much later under the label “computational analysis.” Swade ascribes the enduring myth to Babbage’s friend and advocate Dionysius Lardner, who made a living as a writer and traveling lecturer. Finding that his talk on the mathematical advances of Babbage’s engines was too difficult for mass audiences to understand, he simplified the argument to concentrate solely on the avoidance of errors. When he wrote up his lecture in 1834 it became a standard reference paper and thus, as Swade concludes, “dumbing down has had a defining influence on our historical perception of Babbage’s motives.”
Babbage’s Analytical Engine project is often cited as the first on
OCR for page R18
Electronic Brains: Stories from the Dawn of the Computer Age
which a woman programmer worked. Babbage’s good friend (some say mistress) Ada Byron, Countess of Lovelace, was enchanted by his machines and particularly the mathematics involved and wrote a famous article, “Sketch of the Analytical Engine.” This mixed her translation of a Swiss engineer’s description with her own copious notes, including details of programming examples that most people have attributed to her. However, Doron Swade concludes that Byron’s role in Babbage’s work has been both exaggerated and distorted down the years, and describes her as a precocious novice where mathematics was concerned. Another noted Babbage historian, Bruce Collier, commented tartly, “I guess someone has to be the most overrated figure in the history of computing,” but Swade does not go that far, arguing that Byron at least deserves credit for
her unique understanding and insight into the potential of the computer particularly in areas beyond the confines of mathematics. She wrote of the Analytical Engine manipulating symbols representing entities other than quantity and the extension of the concept of a computer beyond number is not found anywhere else in contemporary commentary, and specifically not in Babbage’s writing. So I would say that Byron deserves to be celebrated as the first person to see beyond Babbage, in a visionary and even prophetic way, the potential of a universal computing machine for application outside calculation.
Byron’s rise to iconic status as the first “programmer” became unstoppable in the mid-twentieth century and was crowned in the 1970s by the naming of the computer language “Ada” in her honor.
Toward the end of the nineteenth century came another
OCR for page R19
Electronic Brains: Stories from the Dawn of the Computer Age
important step on the road to the modern computer. The US Constitution stipulates that a census must be held every ten years in every state for the purpose of calculating the numbers in the House of Representatives. The first census was a simple head count, held in 1790 when the population was under 4 million, and took just nine months. By the late nineteenth century much more information was being gathered about each person, the population was approaching 50 million, and the analysis was taking seven years. Each census took longer than the one before and the government feared it would still be adding up the 1890 census returns by the time the 1900 census took place.
Enter the first revolution in office calculating machines, the mechanical tabulator. This was a set of equipment devised by the inventor Herman Hollerith. It enabled an operator to punch holes in a card corresponding to the data gathered about the individual (one punched card per person). An electrical machine read the card and indicated which of a large number of compartments to place the card in. Using these machines, the human counters could collate and analyze the figures in a fraction of the time it had taken to do everything by hand. The 1890 census results were available in just two years, saving the Census Bureau $5 million (equivalent to $100 million today).
Many companies had similar requirements for analyzing large amounts of data, so this early success started a huge industry in tabulating machines and development was rapid. The 1900 census was tabulated in a mere six weeks, despite a further 50 percent increase in population since 1880. Hollerith’s firm merged with two rivals to become the Computing-Tabulating-Recording Company, and in 1924 it was renamed International Business Machines—soon to be known simply as IBM. Another early rival to Hollerith in the manufacture of tabulating machinery was James Powers, whose company in turn merged with others into the Remington
OCR for page R20
Electronic Brains: Stories from the Dawn of the Computer Age
Rand Corporation in 1927 (Remington had become famous as the maker of the first commercial typewriter, while Rand made the Kardex card index systems). IBM came to dominate the tabulating-machine market, with Remington Rand taking much of what was left, and both companies would have important roles in the computer age.
There was one further period of innovation before the computer age really began. It started with the invention by Vannevar Bush of the Differential Analyzer, which he unveiled in 1930 at the Massachusetts Institute of Technology. Several more Differential Analyzers were built in Britain during the middle years of the decade and most of these were made largely from Meccano, the legendary boy’s construction kit, but they were not toys and were capable of a surprising degree of accuracy. None of these British machines, however, matched Vannevar Bush’s second Analyzer, unveiled in 1935, a 100-ton monster with around 2,000 valves (or “tubes”), almost as many relays, and 150 electric motors. Its calculations were controlled by instructions on paper tape and, although the computations were purely mechanical (the electronic components just controlled the movements of the mechanical ones), it was another significant step, not least for its successful use of so many valves.
The Differential Analyzers were of great importance in the 1930s, enabling the solution in a matter of hours or even minutes of complex differential equations that could take teams of human mathematicians weeks to solve. These were not necessarily obscure mathematical exercises, as differential equations could be used to model weather systems, describe the trajectory of a shell fired from a gun, or calculate the rate of erosion of river banks. They would later find many uses in war-related applications and influence the thinking of several of the early British, American, and Australian computer pioneers.
OCR for page R21
Electronic Brains: Stories from the Dawn of the Computer Age
During the same decade a German engineer, Konrad Zuse, was working on one of the first electromechanical computers, the Z1. Born in 1910, Zuse was still a student when he started thinking about better calculating machines based on three logical principles: program control, the binary system, and floating-point arithmetic.
Zuse later justifiably claimed that “today these concepts are taken for granted but at the time this was new ground for computing.” He started active design in 1934, “working independently and without knowledge of other developments going on around me,” and within two years he finished the “logical plan.” It took two more years to construct his machine before it started working in 1938. His son Horst Zuse claims it was “the first freely programmable binary-based machine in the world.” As usual the carefully framed definition is crucial to the claim, but the Z1 was certainly a remarkable device. It had a memory of 64 words (each word containing 22 “bits,” or “binary digits”). It had a high-performance adder and was capable of floating-point arithmetic, and hence it was able to handle very small or very large numbers with precision. It had a control unit, and the whole machine was programmed by instructions on paper tape. Perhaps the most ingenious parts were the mechanical “yes/no” modules used in both the memory and the arithmetic unit, where Zuse designed a system of sliding pins in a grid of movable thin metal sheets. This was analogous to the modern electronic memory that consists of a large number of bits in a “grid” of rows and columns, where every bit can be individually “addressed” by row number and grid number and then written or read.
Clever though the Z1 was, Zuse found it insufficiently reliable and went on to build the Z2, which used a similar memory but had 800 old telephone relays in the arithmetic unit. That convinced him that electrical relays were reliable enough and he went on to build
OCR for page R22
Electronic Brains: Stories from the Dawn of the Computer Age
the most impressive of the series, the Z3, which used relays throughout. Started in 1939 and completed in 1941, the Z3, Horst Zuse claimed, was “the first reliable, freely programmable, working computer in the world based on a floating-point number and switching system.” It was big, at five meters long, two meters high and almost a meter deep, though not on the scale of the American “Harvard Mark 1” electrical relay computer, which it pre-dated by several years. Konrad Zuse claimed his Z3 could “calculate all mathematical problems” and was even capable of playing chess, though that was never demonstrated. He did, however, develop a sophisticated programming language for the machine, which he called “plankalkul.”
Unfortunately for Zuse, all three machines and most of the drawings were destroyed in air raids that flattened Zuse Apparatebau, the company he had formed in 1940 in Berlin to build them. During the later war years he worked on the Z4, but relentless bombing of the German capital prevented its completion, and in 1945 he fled with the machine to Bavaria, where he hid it in a barn. For two years daily survival amid the shortages of occupied Germany was his main concern, but by 1947 he had the Z4 working again, though mains power was intermittent and he had to fashion spare parts from discarded tin cans. Eventually he persuaded the Swiss Federal Institute of Technology in Zurich to buy the machine and in 1950 he delivered it to the institute, where the scientists were astonished at the mechanical memory (he had reverted to the thin metal sheets and pins of the Z1 and Z2). It was reliable enough to leave running unattended overnight and Zuse once said that “the rattling of the pins and relays was the only interesting thing about Zurich’s nightlife.”
The Z4 did some useful work in Switzerland but, ultimately, half a second to perform a simple addition and six seconds to compute a single division were far too slow to compete with electronic
OCR for page R23
Electronic Brains: Stories from the Dawn of the Computer Age
methods of calculation. It was the end of an era rather than the beginning. Zuse had moreover missed an earlier chance to take the electronic route. Back in 1936 he was lobbied by his friend Helmut Schreyer to use valves which “could switch a million times faster than elements burdened with mechanical and inductive inertia.” However, suitable circuits did not exist then and Schreyer’s own colleagues doubted a machine with thousands of valves would work reliably. Yet before the end of the 1930s American and British pioneers were choosing the electronic option and in a few years would prove the doubters wrong.
OCR for page R24
Electronic Brains: Stories from the Dawn of the Computer Age
This page intentionally left blank.
