Mark I | Research & Encyclopedia Articles

Mark I

The Automatic Sequence Control Calculator (ASCC), also known as Mark I, was the first large American digital computer to work from a program and produce reliable results. It was designed by Howard H. Aiken, a physicist at Harvard University, and built by IBM and the United States Navy.

While in graduate school during the 1930s, Aiken became dissatisfied with calculating differential equations by hand or existing calculators. He decided to build a better machine based on the Analytical Engine of the British computer pioneer Charles Babbage.

Aiken designed a machine to add, subtract, multiply, and divide, calculate exponents, handle probabilities, and look up trigonometric function values on tables. It would read the initial data from a punched card, perform the calculations, and punch the results on other cards. Like Babbage, Aiken organized his data in registers, and he specified fixed decimal points.

In 1937, Aiken contacted several calculating machine manufacturers about building the machine, but they turned him down. Even the president of Harvard discouraged him, saying the idea was impractical. But when Aiken presented his idea to IBM, the company's president, Thomas Watson, agreed to build it.

When construction began, in 1939, vacuum tube (electronic) technology was available. However, the Mark I was an electromechanical machine, making use of the standard parts and engineering expertise available at IBM. In its final form, it was a linkage of seventy-eight adding machines and calculators, over fifty feet long, nine feet high, and weighed five tons. It used over 130 registers to hold data, seventy-two for constants and sixty hand-operated registers for intermediate processing results. Numbers were transferred from one register to another by a combination of mechanical clutches and electrical sensors.

Mark I could perform any specified sequence of operations, for numbers up to twenty-three digits long. Addition or subtraction of two twenty-three-digit numbers took one-third of a second. The program instructions were on punched paper tape. Punched cards, punched tape, or hand-set switches provided the input. The results were punched on cards or printed on an electric typewriter.

Mark I was completed in January 1943, in the midst of World War II. Aiken, who was serving as a Navy officer, moved the machine to Harvard where it performed its first calculations in May 1944. During and after the war, Mark I was used for military purposes, including calculations for ballistics trajectories and the atomic bomb.

When the machine was dedicated in August 1944, Aiken angered IBM's Watson by taking the entire credit for it. As a result, IBM broke its relationship with Aiken and Harvard. But the company combined its Mark I experience with other research as the foundation of its own computer line.

Mark I worked reliably at Harvard for fifteen years. But faster, entirely electronic computers were already being built. From 1945 to 1952, Aiken built three advanced versions of his machine incorporating new technology. Mark II, with electromagnetic relays, and Mark III, with magnetic drum storage, were designed for the Navy. Mark IV, built for the Air Force, used magnetic-core registers.