In 1939, development of the general-purpose electromechanical computer Harvard Mark I began at IBM's laboratories. The company provided money, technicians, and workspace, but the project was directed by Howard Aiken, a Harvard mathematician. Clearly influenced by Babbage's Analytical Engine, the Mark I's guts contained 760,000 gears and more than 800 kilometers of wiring. Capable of solving complex calculations of equations related to the parabolic motion of projectiles, it stored its data and programs on paper tape. It began operating in May 1944.
The Birth of the Mark I
The war raging in Europe demanded all possible effort from mathematicians. The war industry had developed incredible machines and cannons, but complex trajectory calculations were needed to make all that scrap truly useful. The solution was, of course, automation. In 1939, computing was in its infancy, and several places around the world were working against the clock to develop a general-purpose digital computer that worked fast enough and well enough to be useful. There was a great deal of prior theoretical work showing that such a thing was possible, but in practice only a handful of "analog computers" had been built, extremely complex and little more powerful than a calculator.
In 1939, in IBM's laboratories in the United States, development of the Mark I began, which would become the first electromechanical digital computer built in that country. Officially called the Automatic Sequence Controlled Calculator and designed by Harvard mathematician Howard Aiken, the Mark I was what is known as a "general-purpose computer", that is, it could be freely programmed to solve different problems.
Obviously, given the circumstances, its work would essentially be restricted to solving complex calculations related to the parabolic motion of projectiles. Although the machine's development and construction were carried out with IBM's funding and personnel, the design was Aiken's, who had been strongly inspired by the Analytical Engine conceived by Charles Babbage. In 1944, after almost five years of work, the Mark I was ready.
Hardware and Design
The Mark I was physically imposing. It was contained in a glass case measuring about 15.5 meters long, 2.5 meters high, and about 60 centimeters deep. Inside it, some 760,000 wheels and gears, more than 800 kilometers of cables, and thousands of relays were packed. The computer's "guts" were held together by about three million connections.
The whole assembly weighed more than five tons and was equipped with paper tape reader units, which converted the perforations in the medium into electrical signals. These peripherals allowed operators to enter the necessary data and programs relatively quickly, with almost no errors. Once the program was executed, the results were printed by electric typewriters supplied by IBM itself or dumped onto punched cards, the "star medium" of the era.
Despite using a large number of electrical components, many of the Mark I's fundamental parts were mechanical in nature. On the front panel, more than 1,400 rotary switches were lined up, used to display the values of the constant registers that each program needed to run. Part of its memory consisted of 72 mechanical registers, each capable of storing numbers of up to 23 digits. A "0" was used to indicate that a number was positive and a "9" to indicate that it was negative.
The position of the decimal point was fixed when programming each specific task, and could be placed between any two digits. Like Babbage's Analytical Engine, the Automatic Sequence Controlled Calculator could operate in a "double-precision mode", concatenating two 23-digit registers to obtain one of 46 digits. The processing speed, ridiculous compared to modern computers but revolutionary for the time, averaged 5 arithmetic operations (addition, subtraction, multiplication, division, or comparison with a previous result) per second.
Programming and Operation
To solve a problem using the Mark I, operators began by preparing a program—composed of a sequence of instructions—that was encoded by holes (or the absence of them) punched in a long paper tape. Once this process was complete, the tape was fed into the computer, which converted the programmer's holes into instructions and numbers that were transferred from one register to another via electrical signals. At that moment, the machine started working.
The noise it produced was terrible, like hundreds of typewriters working at once. The first version of this computer was not capable of making jumps within the program based on the results obtained. This limitation, which seems inconceivable today, had to be overcome by stopping the machine and replacing the program tape with one containing the specific sequence of instructions needed in each case.
The problem was solved with the addition of a unit called the "Subsidiary Sequence Mechanism", which allowed it to store up to 10 subprograms with a maximum of 22 instructions each. Physically, this addition was a block of three huge circuit boards and three additional tape readers. During the execution of a program, depending on the contents of its registers, the Mark I could "jump" to the contents of any of those tapes.
The Successors
After the war, Aiken continued working on computer development. In 1947, the Mark II was born, and later the Mark III. The latter was the first in the series to incorporate some electronic components. The next and last of the list, the Mark IV, was built entirely with components of this type and used magnetic cores as memory. Furthermore, this machine was capable of storing results on "magnetic drums", devices whose operation was similar to a modern hard disk but which, instead of using the surface of disks as a medium, recorded data on the sides of a cylinder.
A Famous (Possibly False) Quote
Howard Aiken (and often Thomas Watson of IBM) is attributed the phrase "Only six electronic digital computers will be needed to satisfy the computing needs of the United States." However, this statement, theoretically pronounced in 1947, was never actually said by either of them. Regardless, the Mark I laid the foundations for many later computers, and can justly be considered an ancestor of the one you are using to read this.
For more information, see Mark I on Wikipedia.
