April 09, 1998
University Gazette


Full contents
Police Log
Gazette Home
Gazette Archives
News Office



Howard Aiken: Makin' a Computer Wonder

Applied Mathematics Professor Was Computer Pioneer

By Cassie Ferguson

Gazette Staff

The desire for answers to the questions raised by his doctoral thesis in physics led Howard Aiken to the conclusion that he would have to build a calculating machine unlike anything ever seen before at Harvard -- a computer.

Aiken needed numbers for his theory of space-charge conduction in vacuum tubes, but the problems were beyond the capability of desktop calculators of the day. Frustrated by his dilemma, in 1937 he wrote a proposal for a giant calculating machine, one that could represent negative and positive numbers, do standard arithmetic, and carry out more than one operation in a sequence.

"The desire to economize time and mental effort in arithmetical computations, and to eliminate human liability to error is probably as old as the science of arithmetic itself," he wrote, although he would later joke that the computer was "only a lazy man's idea."

A year earlier, in 1936, Aiken had proposed his idea to the Physics Department, which did not see the same need for a computing machine and was reluctant to give up space for one in its building. He was told by the chairman, Frederick Saunders, that a lab technician, Carmelo Lanza, had told him about a similar contraption already stored up in the Science Center attic.

Intrigued, Aiken had Lanza lead him to the machine, which turned out to be a set of brass wheels from English mathematician and philosopher Charles Babbage's unfinished "analytical engine" from nearly 100 years earlier.

Aiken immediately recognized that he and Babbage had the same mechanism in mind. Fortunately for Aiken, where lack of money and poor materials had left Babbage's dream incomplete, he would have much more success.

Later, those brass wheels, along with a set of books that had been given to him by the grandson of Babbage, would occupy a prominent spot in Aiken's office. In an interview with I. Bernard Cohen '37, PhD '47, Victor S. Thomas Professor of the History of Science Emeritus, Aiken pointed to Babbage's books and said, "There's my education in computers, right there; this is the whole thing, everything I took out of a book."

Next fall Cohen has two books on Aiken due to debut from the M.I.T. Press: A Portrait of Howard Aiken, Computing Pioneer and Makin' Numbers: Howard Aiken and the Computer, a collection of essays edited by Cohen and Gregory M. Welch '85.

Plans to Programs

The head of the Physics Department eventually did give in to Aiken's request for space, but Aiken had to build the machine first.

Aiken took his first design to the Monroe Calculating Machine Co., which turned him down, but told him to try IBM's president Thomas J. Watson. He agreed to build Aiken's dream machine for the then outrageous cost of $200,000.

Since IBM funded and build the computer, it wound up consisting of the same mechanical parts the company used to construct its accounting machines, rather than electronics. The first electronic computer, ENIAC, would be built a few years later at the University of Pennsylvania soon after Aiken's machine in 1946.

Construction of the computer started in 1937 and continued through the end of 1943. Robert V. D. Campbell, MA '48, supervised the final assembly of the machine in an IBM plant in Endicott, N.Y.

The finished product stood 8 feet high, 51 feet long, and 2 feet wide. Although the machine might not have been the first electromechanical computer to be built, many computer pioneers believed that it sparked the computer age. The computer weighed five tons and consisted of about 760,000 parts, including 2,200 counter wheels, 3,300 relay components, and 530 miles of wire.

To work the machine, a person had to write a program converting problems into a code that could be read by the computer. That code was then converted into a series of holes punched into a paper roll of tape, each representing a single instruction. After being inserted into a tape reader, a series of feelers would find the holes, closing a relay switch every time one was found. Those relay switches routed information to other parts of the machine where numbers were stored in registers.

Counters, mechanical tables, and sensing circuits performed their calculations based on the numbers stored in those registers and the end results were printed by a set of automated typewriters.

Frequently-used sets of instructions could be stored for use in future problems, saving the time it would take to reprogram them. Grace Hopper, who worked for Aiken and who later invented the programming language COBOL, pioneered those routines. Programmers now call them library functions. She also claimed that she found the first computer "bug" - a moth crushed on a relay switch.

Computing Victory

Aiken's computer, originally named the IBM Automatic Sequence Controlled Calculator and later the Harvard Mark I, ran at the pace of three calculations per second, a turtle compared to today's simplest digital calculators.

In 1944 the speed was considered unbelievably fast. According to a New York Times article, "At the dictation of a mathematician, it will solve in a matter of hours equations never before solved because of their intricacy and the enormous time and personnel which would be required to work them out on ordinary office calculators."

The first two problems the computer tackled were from physics and astronomy: calculating integrals and producing numbers to be used to design a telephoto lens.

Later the computer worked on problems associated with magnetic fields, radar, and a top secret equation from scientists at Los Alamos Laboratories, N.M., concerning the implosion of the atomic bomb.

While the machine was running, the ticking sound of thousands of registers turning filled the Physics Department basement. "It sounded like a clickety-clackety rhythm band," said Anthony Oettinger '51, PhD '54, Gordon McKay Professor of Applied Mathematics, who was one of Aiken's graduate students.

The giant machine ran 24 hours a day. Whenever it stopped, a bell would ring, alerting one of the people tending it to press a button or turn a knob to prompt the computer to move on to the next step. Often Aiken would pop out of his nearby office, which was filled with awards, to see what was going on.

On Aug. 14, 1944, the University formally dedicated the computer and it continued to run for 14 more years. Aiken was involved in the construction of three more computers, as well as establishing at Harvard the world's first full-scale degree program in what we now call computer science.

Now, more than 50 years later, part of the Mark I sits in the lobby of the Science Center, another section is in the Smithsonian Museum of American History, and the last part is in IBM's historical collection.


Copyright 1998 President and Fellows of Harvard College