PART 3: BIG GUNS AND BIG COMPUTERS
"I think there is a world market for maybe five computers."
--- Thomas Watts, chairman of IBM, 1943
We owe much of the development of the modern electronic computer to World War II. There were codes to crack, ballistics to calculate, and a lot of other problems that required hours of tedious number-crunching. The faster these tasks could be accomplished, the sooner the war could be won, and so the governments of the US, Britain, and Germany began to pour millions of dollars into computer research and development. In America, one of the first computers developed for the military was the differential analyzer constructed in 1940 at The University of Pennsylvania's Moore School. It was used to calculate artillery ballistics tables. These calculations were nothing to take lightly - figuring out the optimum shape of an artillery shell could take over 20,000,000 separate calculations. While these early computers were incredibly slow by today's standards - each operation could take 3 or 4 seconds, compared to modern home computers that can perform millions of calculations per second - they could do it 24 hours a day, day in and day out, and never make a mistake.
Meanwhile, in over in Britain, Alan Turing was working as a cryptographer, decoding German codes. One of his triumphs was breaking Germany's famous ENIGMA code. But the Germans had an even more complicated code for their top-secret communications that was generated by a machine they called Geheimfernschrieber or secret telegraph. The British just called it he "Fish".
In January, 1943, Turing began to work on an electronic computer to try to break the Fish. By the end of the year they had it completed. It used more than 1500 vacuum tubes and was called the COLOSSUS <<PIC>> . It was a special purpose machine that was unable to do even simple decimal multiplication, but it cracked the codes, and the lessons they learned in building it served to further the development of computers after the war.
The wartime efforts of scientists and engineers were also making significant advancements for the guys on the other side. It was not until the end of W.W.II that the allies discovered the amazing accomplishments of a German construction engineer named Konrad Zuse. In 1941 he had built a program controlled calculator called the Z3, three years before the Americans achieved a similar milestone. Zuse was a true geek, assembling his first computer, the Z1, in his parents living room in Berlin in 1938. Wartime shortages forced him to be creative in his innovations. He used relays instead of vacuum tubes which were in short supply. Paper was also scarce so he was unable to use punched cards to store the data. In a clever stroke of geek innovation, he punched holes in old movie films for his programs. His achievements are doubly amazing because Zuse was unaware of the work that had been done by other early computer experimenters and pioneered his innovations on a clean slate.
The Z3 was a very advanced machine for the time. It utilized the binary number system and was capable of sophisticated operations like floating-point arithmetic. Unfortunately, his Z3 was destroyed by Allied bombing, but a later model, the Z4, was hidden in a cave in the Alps and survived to be used in the fifties by a Swiss bank. In a happy retrospective note, the Z3 was reconstructed by admiring retro-geeks in the sixties. Zuse continued his work with computers after the war, continuing to come up with ideas that were ahead of his time. In 1958 he theorized the concept of parallel computing years before it became a reality.
Serious geeks generally agree that the first large-scale digital computer was the Harvard Mark I <<PIC>> which was developed over a five year period between 1939 and 1944 by Howard H. Aiken and his colleagues at IBM. Originally called the IBM Automatic Sequence Controlled Calculator (ASCC), it was the first of a series of computers created under his direction and funded by IBM and the US Navy. Since punched cards had been in common use by IBM for many years, Aiken based his machine on existing parts. By the time it was finished, it was a 5 ton electro-mechanical monster that was 50 feet long, 8 feet high, and contained nearly a million switches, relays and gears. One contemporary observer described it as sounding like a room full of ladies knitting. It may have been slow and noisy, but it could handle complex mathematical operations and kept on ticking until it was finally taken out of service in 1959. It's now on display at the Harvard Science Center.
Here's an interesting and useful bit of trivia involving the Mark I's successor, the Mark II. The first recorded use of the word "bug" to describe a computer problem happened on September 9th, 1945, at the Naval Weapons Center in Dahlgren, Virginia, when a moth flew into one of the Mark II's relays and jammed it. When they extracted the culprit, they taped it into their logbook along with the official report stating. "First actual case of a bug being found". Legend has it that Grace Murray Hopper, a naval officer and famous mathematician was the one who found the bug, but she has stated categorically that she wasn't there at the time. Incidentally, in a further bit of trivia, Grace Hopper was the first woman to achieve the rank of Rear Admiral in the US Navy (1983). She's also one of the pioneers in the field of data processing and was head of the project that developed the COBOL programming language.
Now let's take a look at the famous ENIAC (Electrical Numerical Integrator and Calculator) <<PIC>> , the first computer to capture the public imagination. It was developed between 1943 and 1946 by John W. Mauchly and J. Presper Eckert Jr. at the Moore School of Engineering at the University of Pennsylvania. If the Harvard Mark I was a monster, the ENIAC was the Godzilla of computers. It was a 30 ton behemoth, 100 feet long and 8 feet high, with 18.000 vacuum tubes, 3600 dials, 1500 relays, 70,000 thousand resistors and 10,000 capacitors. All this hardware was housed in forty-two cabinets that were two feet wide and nine feet tall, arranged in a U shape. The ENIAC and the machinery required to keep those 18,000 vacuum tubes cool took up 1800 square feet (167 sq. meters) of floor space and sucked up over 170 kilowatts of electricity. It had less computing power than today's $5.00 K-Mart calculator.
It used punched cards to input and output information, and to make it do anything, the engineers had to instruct it by plugging in hundreds of cables and adjusting the 3600 dials. But it was a start. It was faster and more accurate than pencil and paper or previous mechanical calculators, and the future of computing looked bright. In March, 1949, Popular Mechanics magazine featured an article on the ENIAC and offered this optimistic prediction: "Where a computer like the ENIAC is equipped with 18,000 vacuum tubes and weighs 30 tons, computers in the future may have only 1,000 vacuum tubes and weigh only 1 1/2 tons." Most of the ENIAC's down-time was the result of locating and replacing burned out tubes, and with 18,000 tubes, it was a needle-in-a-haystack nightmare. In 1952 alone there were nearly 20,000 tube failures - which works out to about 50 tubes a day.
The ENIAC was unveiled for the public on Valentine's Day, 1946. The following month, its designers, Eckert and Mauchly, left the University of Pennsylvania and started their own computer manufacturing business, the Electronic Control Company.
Meanwhile, back at the ranch...
Mathematician John vonNeumann was following the development of the ENIAC and noticed a glaring omission in its design. It couldn't store programs or data, but had to rely on punched cards to hold this information. He thought about it for awhile and came up with the idea of the "stored program", the combination of CPU and separate memory storage that are the basis for the architecture of most modern computers. His concept became known as - yep, you guessed it - the Von Neumann Model. What all this meant was that, rather than being implemented anew each time the computer was programmed for a particular project, parts of the program, called subroutines, could be stored and recalled when they were needed.
Eckert and Mauchly had been working on a similar ideas and had designed a machine they called, with the early geek fascination for acronyms, the EDVAC - the Electronic Discrete Variable Arithmetic Computer. They finished the design shortly before they abandoned the world of academia for the world of business, but their successors didn't complete it until 1952. Many geeks consider the EDSAC to be the first stored-program computer, but a small, experimental machine based on von Neumann's ideas was built in England at Manchester University in 1948.
A similar machine called the EDSAC (Electronic Delay Storage Automatic Computer) was developed by Maurice V. Wilkes at Cambridge University in England, and was up and running in May, 1949. The EDVAC and the EDSAC were the first of a new generation of computers to use random access memory (RAM). Granted, this new RAM wasn't very big - it could only hold the equivalent of about a thousand words - but it was ample for the programs of the time, and it paved the way for the development of a new commercial computer industry. These new machines were faster, more efficient and certainly more compact than the ENIAC - they weren't much bigger than a large desk and operated with no more that a couple of thousand tubes.
In 1949, Eckert and Mauchly's new company built one of these new machines, called the BINAC for Northrup Aviation. The following year, after only four years of operation, the Eckert-Mauchly Computer Corporation was sold to Remington-Rand, because of financial difficulties that stemmed from contract problems during the notorious McCarthy era.
A duplicate of the EDSAC was one of the first two computers to be used for commercial data processing, and Remington-Rand, with the technology they acquired from Eckert and Mauchly, built the other. Their Univac I, following the tradition of Hollerith's Tabulating Machine sixty years earlier, was built for the 1950 US Census. It was also used to attempt to predict the outcome of the 1952 presidential elections. It gave Eisenhower a landslide victory over Adlai Stevenson, but the programmers thought it was in error and re-adjusted the variables for a more modest victory. In the end, Eisenhower won by almost the exact margin of UNIVAC's original prediction.
There were other computers built in the early fifties - the Whirlwind at MIT in 1951, the IAN at Princeton's Institute for Advanced Study, the EDVAC, the ORDVAC, the MANIAC (my personal favorite), and IBM's 701, their first commercial stored-program computer - but the really big event was the invention of the transistor. It signaled the end of the era of vacuum tube computers (as well as radios and other consumer electronics) and the dawn of the modern computer.