ORIGINAL COMPUTERS: "Computer" Alice Burks and her husband, ENIAC design team member Arthur Burks, stand behind an accumulator unit of the machine.
"Eniac holds a place in the history books as the first general-purpose electronic computer that could be programmed to perform different types of calculations at high rates of speed," notes Arthur Burks, a computer scientist who was part of the ENIAC team. He is now an emeritus professor in the department of computer science--which he helped found--at the University of Michigan.
As part of the commemoration, Jan Van der Spiegel, a professor of electrical engineering at Penn, and a group of his students are attempting to "rebuild the whole computer on a single chip. We will be using modern-day technology to make a chip with the internal architecture and logical structure that mimics the capabilities of the original machine," he explains.
Another group, headed by electrical engineering professor Fred Ketterer, is working on reconstructing a part of ENIAC "to show how the original machine functioned," adds Van der Spiegel, who is also involved in that project. "Although the unit will be much smaller, the unit will have circuitry of the original and operate in the same way."
Built by a team of civilian and Army scientists at the Moore School of Engineering, the completed computer consisted of 18,000 vacuum tubes, 70,000 resistors, and 10,000 capacitors, arranged in 40 panels in an-80-foot "U." It weighed about 30 tons, and took up a whole room.
Today, relics are all that remain of that massive machine, and constitute a permanent exhibit in the same room at Penn. A large, black, cabinet-like panel with a numerical display unit resembling several cash registers piled one upon the other bears the legend ENIAC. A metal box--called an accumulator--with several columns of knobs and switches lies to one side of the cabinet. On a nearby table are a few troughs with rows of glass vacuum tubes sticking out along their length. For the scientists who built ENIAC, the modest collection evokes vivid memories.
ENIAC represents one of the first major steps in the computer revolution, say those veteran scientists, who participated in designing and developing this and other early computers. Back in those early days of World War II, reminisces Herman Goldstine, a mathematician and computer scientist who was also part of the team that built ENIAC, the very word "computer" had a different meaning; it stood for a person--not a machine--who did calculations.
"During the war, a large number of women were hired as `computers,' to perform the calculations necessary for preparing firing and bombing tables," recounts Goldstine, who is currently the executive officer of the Philadelphia-based American Philosophical Society. These were extensive tables accompanying each weapon that enabled artillery operators to position their guns properly with respect to the distance and angle of the target, taking into account how numerous other factors--including wind speed, the weights of the shells, and propellant charges--affected the paths of the projectiles. According to Goldstine, then a first lieutenant in the army assigned to the Ballistic Research Laboratory at the Aberdeen Proving Grounds in Maryland, it was to automate this process that the War Department first decided to look into building an electronic computer.
"There was a need to calculate these tables for every possible combination of gun, shell, and fuse. The men were all drafted to go to war, and pretty soon Aberdeen ran out of girls to do the job," he quips.
"The truth is that we couldn't do the calculations fast enough," offers Alice Burks, Arthur Burks's wife, who had worked as a computer at Penn's Moore School of Engineering in Philadelphia during that time, and has coauthored with her husband numerous writings about ENIAC and the history of computers.
"There were about 75 women working at the Moore school," she recalls, "and, using mechanical desk calculators, it took us about one or two days to do a series of calculations. We needed a quicker way."
This search for a quicker way led to a collaboration between scientists at the Moore School and the Ballistic Research Laboratory to build ENIAC. The effort was planned and headed by two scientists from Penn--John W. Mauchly, a physicist, and J. Presper Eckert, Jr., an expert in electronics. The Army first granted funding for the project in April 1943.
By December 1945, says Arthur Burks, a working computer was ready: "ENIAC solved its first problem--calculations for the hydrogen bomb--in December, but the machine was not complete in every detail yet." Two months later, on Feb. 14, 1946, the computer was unveiled to the public at the Moore Schoo. Arthur Burks gave the demonstration. "We showed the machine's ability to add 5,000 numbers in one second," he remarks. "Just before I flipped the switch I warned the journalists to look up from their notebooks," so as to catch the numbers changing. "But most of the people were still writing when the results appeared."
Certainly the machine was a vast improvement over the human computers, conducting the various operations that normally took them one to two days in about 20 seconds, Arthur Burks remembers.
Goldstine points to a crucial advance during this period that allowed Mauchly, Eckert, and others to start thinking about the feasibility of ENIAC. This was "the introduction of vacuum tubes, which enormously sped up the computation process, which until then had been performed by mechanical relays in earlier computers like the punch-card machines."
In fact, the Burkses point out, ENIAC was not the first digital computing machine. Between 1935 and 1942, a scientist named John Atanasoff "had designed an electronic computer based on vacuum tubes, for the specific purpose of solving simultaneous linear equations" at Iowa State College (now Iowa State University), says Arthur Burks. Mauchly had visited Atanasoff's home in Ames, and observed the Atanasoff-Berry computer (ABC) before he and Eckert had begun working on ENIAC.
This visit, and subsequent correspondence between Mauchly and Atanasoff, "had an important influence on Mauchly's ideas for designing ENIAC," agrees Goldstine.
Despite its advantages--in 1946 the Philadelphia Inquirer reported that it "wipes out the boundaries imposed by time on the limits of mortal thinking"--ENIAC had some serious drawbacks, which Goldstine and others set about to rectify in their subsequent trials.
"The main difficulty was that this was not a stored program computer. Once we set it up solve a problem it could do so very quickly. But every time we needed to change it we had to go back and reprogram it, which would take up to two days," he states. A program was set by manually plugging in cables and setting the switches of the different components of the computer.
The next step, undertaken independently by different members of the ENIAC team, which had split up after the wartime effort, was to make stored-program computers. One such machine was designed under the leadership of Goldstine and computer science pioneer John von Neumann at the Institute for Advanced Study in Princeton. Mauchly and Eckert, who had been fired from the University of Pennsylvania a year after launching ENIAC as a result of disputes over rights to the machine, began their own independent company, called the Electronic Control Co. This company went on to build UNIVAC (for Universal Automatic Computer) for the U.S. Census Bureau.
EXPERIENCING EVOLUTION: Computer pioneer Herman Goldsine, who helped design ENIAC, marvels at how much computers have advanced in the 50 years since they came into being.
"ENIAC bears about as much resemblance to a modern computer as dinosaurs do to humans," comments Goldstine. Certainly, none of the inventors of early computers had any idea of the outcome of their creation, he professes. "Our vision then was for a modest number of computers in some of the universities and large industries," he remembers. "We hadn't yet thought of the world being taken over by them."
"We all understood and believed [the computer] would revolutionize science and engineering," adds Arthur Burks. But he agrees that he and the others who conceived of this machine had never envisioned the extent of the revolution or the breadth of its influence on the general population.
"Thinking about ENIAC gives a warm feeling because it was the first, but it's amazing to see how far we've come," says Marcy Rosenkrantz, associate director for supercomputing technology at the Cornell Theory Center, which houses one of the National Science Foundation's nine supercomputing facilities. "Nowadays we have computers with immense capabilities on our desks, or carry them in our bags. Back then, the ENIAC occupied a whole room and had a host of people with vacuum tubes to run it."
"What is really interesting is how much technology has advanced in such a short time and how much smaller everything is now," concurs Van der Spiegel.
Looking back, Alice Burks declares, "We knew [computers] would shrink, but not how much!" Over the next 50 years, she predicts: "They will still be getting smaller, faster, and cheaper.
"That aspect doesn't change--only the rate seems to slow down."
- (The Scientist, Vol:9, #16, pg.3 , August 21, 1995)
- (Copyright © The Scientist, Inc.)