Michigan autocoder which included differential operations,

University of Michigan 1955

  • Martin, E. Wayne Jr.; Hall, Dale J. "Data Processing: Automation in Calculation" Review of Educational Research, Vol. 30, No. 5, The Methodology of Educational Research (Dec., 1960), 522-535. view details Abstract: Availability of the electronic computer makes it possible currently to
    employ new methods in many areas of research. Performance of 1 million
    multiplications on a desk calculator is estimated to require about five vears
    and to cost $25,000. On an early scientific computer, a million
    multiplications required eight minutes and cost (exclusive of programing
    and input preparation) about $10. With the recent LARC computer,
    1 million multiplications require eight seconds and cost about
    50 cents (Householder, 1956). Obviously it is imperative that researchers
    examine their methods in light of the abilities of the computer.
    It should be noted that much of the information published on computers
    and their use has not appeared in educational or psychological literature
    but rather in publications specifically concerned with computers. mathematics,
    engineering, and business. The following selective survey is intended
    to guide the beginner into this broad and sometimes confusing
    area. It is not an exhaustive survey. It is presumed that the reader has
    access to the excellent Wrigley (29571 article; so the major purpose of
    this review is to note additions since 1957.
    The following topics are discussed: equipment availabilitv, knowledge
    needed to use computers, general references, programing the computer,
    numerical analysis, statistical techniques, operations research, and mechanization
    of thought processes. Extract: Interpretive Systems
    Interpretive Systems
    Among the first approaches to automatic programing were the interpretive
    systems, in which the pseudo instructions of the programing language
    were stored in the memory of the computer, along with a program
    that translated these pseudo instructions into the proper sequence of
    machine instructions as the computer engaged in the process of solution.
    The most widely known general-purpose interpretive system for the
    IBM 650 is "Bell Telephone Laboratories Interpretive Code," which has
    been described by Wolontis (1956), Andree (1958), and Wrubel (1959).
    Other systems were developed for special purposes, such as the University
    of Michigan "MITLAC" (1955), which included differential equation
    operations, and "SIS" (Haynam, 1957), which is designed for the solution
    of routine statistical problems. Frequently one of the existing interpretive
    systems will lend itself to the solution of the problem at hand;
    however, since time is required to perform the translation of each program
    run, this convenience must be paid for in terms of computer execution