Harvard-Ibm Automatic Sequence Controlled Calculator (Mark I)
The Harvard-IBM Automatic Sequence Controlled Calculator was one of the first general-purpose, digital computers. The computer was also known as the "Mark I." Howard H. Aiken (1900-1973) was the progenitor of the Mark I. While working on his doctoral thesis in physics, Aiken encountered calculations that were long and tedious to complete. He began to entertain the idea of constructing a machine that, once programmed, would automatically perform a wide range of complicated calculations. In 1937, Aiken wrote a proposal to construct a computer, but neither Harvard University, where he worked, nor a company that he initially approached for funding, showed much enthusiasm. However, International Business Machines (IBM) Corporation did pay attention to Aiken's proposal, and agreed to fund the project. IBM management looked on the Mark I project as an opportunity to be associated both with a prestigious university, as well as a "hi-tech" project; they had no plans to commercially develop the Mark I.
The Mark I was constructed at IBM's Endicott, New York facility. Aiken worked closely with IBM engineers to ensure that the new computer would meet his performance goals. The entrance of the United States into World War II limited the resources that IBM could invest in the project. In spite of numerous obstacles, the Mark I was operational by Christmas of 1943. It was dismantled and then installed at the Harvard campus in Cambridge, Massachusetts in March 1944.
The completed Mark I was a large machine, with a height of eight feet and a length of approximately 50 feet. The computer incorporated around three-quarters of a million individual parts. Mark I performed its computations mechanically, through the opening and closing of metal relays, which in turn controlled the flow of electricity in various circuits. The relays and other components were actuated by shafts, which in turn were powered by electric motors. The technology was conceptually similar to the Bell Laboratories relay computers that were functional before and during the operational lifetime of the Mark I.
Input to the Mark I was delivered on punched paper tapes. The punches in the tape represented information. Machines would translate that information into electrical signals for processing. Originally, there were four tape-reading machines, one to input instructions (i.e., programming) and the other three for data. The basic clock cycle of the Mark I was 0.3 seconds, which was synchronized to match the time it took the program tape to advance from one instruction to the next. Output from the Mark I was typed onto paper via automated electric typewriters.
Calculations within the Mark I were carried out decimally (i.e., in base-10 notation), in a fixed-point architecture. The computer word length was 23 decimal-digits, plus a 24th place reserved to indicate sign (i.e., positive or negative). While data was input by paper tape, there were also 60 registers (i.e., memory units) that were set manually to represent particular mathematical "constants." Each of the sixty registers used 24 dials, one dial to represent each of the 23 decimal digits and the sign. These 60 registers formed a kind of manual read-only memory. The Mark I used 72 "accumulators" to temporally store numbers for processing. There were separate units that accessed the accumulators to perform addition, subtraction, multiplication and division. An addition took one clock cycle (0.3 second), while a multiplication required approximately 20 clock cycles, or six seconds.
Originally, conditional branching (e.g., an "if-then" statement) was not a programming option. However, in later years the Mark I underwent several hardware changes to allow primitive branching. One such change was the addition of several paper tapes and paper tape readers to input a program. One such change was the addition of several paper tapes and paper tape readers to input a program; the transfer from one paper tape to another formed the conditional branch.
The Mark I was historically important in at least two respects: during its operational lifetime it carried out many useful and important computations, most notably calculations relating to the design of the first atomic bombs; secondly, because the Mark I was so well known to scientists, government and military leaders, and the public at large, its existence helped condition people to the notion that large computers were invaluable aids to technological and scientific progress. However, the Mark I's use of mechanical action to perform calculations was a technological dead-end. The completion of ENIAC (Electronic Numerical Integrator And Computer) in late 1945 represented the future of computing. ENIAC was a general-purpose computer that employed electronics (in the form of vacuum tubes) to perform its computations, and it could perform calculations much faster than could the Mark I (for instance, about 3 additions per second for the Mark I, versus 5,000 additions per second for ENIAC). By the mid-forties, Aiken had embraced the idea of electronic computers, and he went on to create new computers at Harvard that used electronic components. For convenience, the Harvard-IBM Automatic Sequence Controlled Calculator was called the "Mark I," and subsequent computers from Aiken were called the Mark II and Mark III.
This is the complete article, containing 838 words
(approx. 3 pages at 300 words per page).
