href="http://www-03.ibm.com/ibm/history/exhibits/markI/markI_feeds.html" target="_blank">Feeds, Speeds, and Specifications: ASCC Major Units has a detailed description of the machine from left to right:
"The ASCC consisted of seven major units. From left to right (as one faced the machine), they were:
"Two sections for 60 constants: the dial switches for setting up known values for use as required in the computation of problems.
"Seven sections containing 72 storage counters: Used for storing intermediate results utilized in mathematical processes up to the machine's capacity of 23 digits, and also used in addition, subtraction and accumulation.
"Three sections containing the multiplying-dividing unit: Used a new system of multiplying and dividing through use of the nine multiples of the multiplicand and divisor. It multiplied and divided to the full capacity of the machine of 23 digits.
"Two sections of functional counters: Controlled interpolation of functions and computed logarithms, anti-logarithms, trigonometric functions and printing.
"Three sections of interpolators: tape feeding mechanisms which automatically selected values required in the interpolation process.
"One section for sequence control: Directed the machine's operation in the solution of a problem.
"And at the very right-end of the machine:
"One section containing the Electromatic Typewriters, card feeds and summary card punch. The typewriters recorded the final solution of the problem in printed form. The card feeds automatically fed into the ASCC various data necessary to solve the problem. And the card punch automatically recoded data pertinent to the solution."
The IBM Automatic Sequence Controlled Calculator (referred to as the IBM ASCC but better known as the Harvard Mark I ) was the first operating machine that could execute long computations automatically. Dr. Howard Aiken of Harvard University conceived of and led the project, which was built by IBM engineers in Endicott, New York, before being shipped to Cambridge, Massachusetts in February 1944. The principals at IBM included senior engineers Clair D. Lake, Benjamin Durfee, and Frank Hamilton, whom Aiken considered his co-inventors, plus James W. Bryce, IBM's chief engineer, and Thomas J. Watson, Sr., IBM's president.
The massive instrument had a steel frame 51 feet long and 8 feet high to hold the calculator, which consisted of an interlocking panel of small gears, counters, switches and control circuits, all only a few inches in depth. The Mark I used 500 miles of wire with 3 million connections, 3,500 multipole relays with 35,000 contacts, 2,225 counters, 1464 tenpole switches, and tiers of 72 adding machines, each with 23 significant numbers. It was the industry's largest electromechanical calculator.
How did it work? According to IBM, "The Mark I was a parallel synchronous calculator that could perform table lookup and the four fundamental arithmetic operations, in any specified sequence, on numbers up to 23 decimal digits in length. It had 60 switch registers for constants, 72 storage counters for intermediate results, a central multiplying-dividing unit, functional counters for computing transcendental functions, and three interpolators for reading functions punched into perforated tape. Numerical input was in the form of punched cards, paper tape or manually set switches. The output was printed by electric typewriters or punched into cards. Sequencing of operations was accomplished by a perforated tape."
The operator did not need to be a trained mathematician, but the problem to be solved needed to be programmed by a mathematician using a code book. The code holes were punched in a paper tape, which was then fed into the machine. Tapes coded with useful functions were added to a tape library for future use. The code book, which tried to include every known type of mathematical problem, was written by Howard H. Aiken, assisted by R. V. D. Campbell.
Howard Aiken first conceived of the machine to solve nonlinear equations when he was a graduate student in physics in 1937. Harlow Shapley, Director of the Harvard College Observatory, and T. H. Brown, Harvard Professor of Business Statistics recommended IBM as a company best suited to build such a machine. Aiken approached James Bryce at IBM that year. The project was formally approved by IBM's directors in 1939.
The machine was built and tested in IBM's Endicott, NY laboratory. It was shipped to Harvard on 31 January 1944, and arrived the next day. Lake, Durfee, and D. R. Piatt of IBM supervised the assembly, testing, and modification of the machine during February 1944 in the Cruft Research Laboratory of Physics at Harvard University. By March 15th, the Mark I was in operational condition.
The computer was officially presented by IBM's president Thomas J. Watson to Harvard's president James B. Conant in a ceremony on 7 August 1944. By then, IBM had spent approximately $200,000 on the project and donated an additional $100,000 to Harvard to cover the Mark I's operating expenses.
At first, the computer was used exclusively by the U.S. Navy Bureau of Ships during World War II to run repetitive calculations for the production of mathematical tables. Problems solved by the Mark I include the tracing of light rays in the design of telephoto lenses for the Air Force; the mapping of magnetic fields associated with the protection of ships from magnetic mines; and the production of tables of Bessel functions needed by the Naval Research Laboratory for the design and use of radar. From Los Alamos, John von Neumann brought a set of calculations for implosions. Richard M. Bloch, a senior analyst and programmer for Mark I, handled this problem. Only a year or so later did Aiken, Bloch, and others at the Computation Laboratory learn that these calculations were required for the design of the atomic bomb.
The machine ran twenty-four hours a day, seven days a week.
The work was supported by a technical staff, which included former IBM employees serving in the Navy. After the war, the Mark I operated at Harvard for about fifteen more years solving complex mathematical problems in a number of disciplines.
After its use was discontinued, a portion of the computer was preserved at Harvard as an exhibit. Another portion was sent to the Smithsonian Institution in Washington, D.C., and some of the ASCC's electromechanical counters were preserved in IBM's collection of historical computing devices. The Mark I at Harvard is now part of the Collection of Historical Scientific Instruments.
For a detailed description of the computer and its production history, please see the website maintained by the IBM corporate archives here
The Collection of Historical Scientific Instruments has a scrapbook of photographs dating from 1944, which show the MARK I being assembled in the Cruft Laboratory of the Harvard Physics Department and take the story up through the August dedication ceremony. (See Lib.1964, the full scrapbook, and Lib.1964-001 through Lib.1964.109, for each individual page.)