HPS282S - HISTORY OF TECHNOLOGY AND ENGINEERING
Spring 1998
Guns, Bureaucrats, and Computers
2 April 1998
I. Some Questions
A. Is there a computer revolution?
B. What is the computer revolution?
C. Who/what made the computer revolution?
II. Strands of Development
A. Mathematics and Applied Mathematics
1. Mechanical Mathematicians and Theoreticians
a) Blaise Pascal (1623-62)) makes a mechanical calculator in 1642. Elementary addition and subtraction.
b) Gottfried Wilhelm von Leibniz (1646-1716) in 1671. Besides addition and subtraction could do multiplication by multiple additions.
c) Charles Babbage (1792-1871) set on building a "Difference Engine" in 1820s. Mechanical device never finished.
(1) "Analytical Engine" that had many of the features of modern computers. Also never finished.
2. Theoreticians
a) George Boole (1815-1864) writes An Investigation of the Laws of Thought, on which are founded the Mathematical Theories of Logic and Probabilities (1854).
b) Claude Shannon (1916- ), founder of information theory.
c) Alan Turing (1912-54) describes a theoretical machine (1937) "Turing Machine" that operates on a moving tape with a given alphabet of symbols and the kinds of rules for the relations between the symbols and the successive internal states of the machine.
B. Materials Science and Electronics
1. The Vacuum Tube
a) Comes out of the incandescent light bulb.
b) John Ambrose Fleming (1849-1945), who worked for Marconi, invents the diode, which acts as a kind of electronic valve
c) Later on electronic vacuum tubes are used in radio and telephony as receivers and amplifiers.
d) Much research on this devices done in Bell Labs and GE R&D Labs.
2. Ferdinand Braun (1850-1918) shares Nobel in Physics with Marconi in 1909 for his work on the behaviour of crystals.
a) Current flowing in crystals sometimes flows in one direction (can thus be rectified) and the amount of current flowing does not follow Ohm's Law. Resistance not constant but varies with the amount of current.
b) Braun's work looked down upon by the more prestigious theoretical physicists of the German academic community. But eventually becomes the basis for solid state physics thirty years after his death (Mid 20c)
C. Bureaucratic and Statistical Problems
1. The Corporate Context
a) The growth of large corporations with large numbers of employees, customers, accounts, and administrative records leads to serious strains on traditional methods of taking, keeping, and analyzing records. Some have argued that this demand-pull would have inevitably led to the birth of the computer. Some mechanical and more efficient way was necessary for handling business information. Otherwise the businesses would be threatened with paralysis and administration would choke in red tape and drown in paper.
b) Growth of business appliance industry: typewriters (Christopher Lathan Sholes (1819-90) in 1870s), various kinds of tabulating machines and mechanical calculators (Wm. S. Burroughs (1855-98) first successful recording adding machine in 1888), cash register developed in the 1870s and 80s in Dayton, Ohio at National Cash Register Company.
2. United States Census
a) 1890 census takes one-third of time to tabulate results as 1880 census in spite of 25% growth in population thanks to Herman Hollerith (1860-1929), statistician who uses punched cards
(1) Hollerith leaves to set up his own business for machines that could count and tabulate punched cards using electromechanical means. His company (Computing-Tabulating-Recording Company) eventually became IBM in 1924.
(2) His successor James Powers, also leaves, sets up a company that eventually merged with Remington Rand (rifles, typewriters) that eventually formed (1955) Sperry Rand which was a major manufacturer of computers in early years after WW2.
III. Milestones of Development
A. Analog and Digital Computers in the interwar period
1. Fairly intense activity at a number of sites (universities, industrial r&d labs, and occasionally individuals such as Zuse before he gained any institutional support). Only the most notable are mentioned here. None attracted really serious attention outside academic and engineering circles.
2. Vannevar Bush at MIT develops analog computer. Interest in these "differential analyzers" due to the problems of network analysis for electrical power grids.
3. Konrad Zuse (1910 - ) Originally develops a mechanical computer using binary number system and during the war comes up with various electromechanical versions. The first computer in the world with automatic control of its operations to really work. Done in complete ignorance of other work in US. In 1950 the Zuse Z-4 computer was the only operational computer in Europe. By that time better computers had been built in America and Great Britain
4. Work on calculating machines also being done at Bell Laboratories and Harvard where the Mark I by Howard Aiken (1900-73), with IBM support, an electromechanical computer begins to work in 1944. Culmination of program of research that had begun before the war. Also known as IBM ASCC (Automatic Sequence Controlled Calculator)
B. World War II
1. ENIAC (Electronic Numerical Integrator and Calculator) developed at Moore School of Electrical Engineering University of Pennsylvania by John Presper Eckert (1919 - ) and John W. Mauchly (1907-80). Uses vacuum tubes. Enormous contraption: 18,000 vacuum tubes, weighs 30T, 8' high, 3' wide, almost 100' long, consumed 140 kw power and cost just under $0.5M. 100 x larger than largest piece of electronic equipment built before that time. Quantum jump in size for electronic device. Computes 1000x faster than electromechanical devices.
a) Writing in 1977 [Sci. Am. quoted in Forester], Robert Noyce, founder of Intel, wrote that a microcomputer costing $300 is 20x faster, has a larger memory, is thousands of times more reliable, consumes the power of a light bulb rather than a locomotive, occupies 1/30,000 the volume and costs 1/10,000 as much. You can get it mass produced in your corner computer store.
b) Put another way, Intel's first microprocessor chip (4004, which holds 2250 microminiaturized transistors on a chip less than 1/8" wide and 1/6" long has the computing power of an ENIAC)
c) Built from 1943 to 1946 to calculate ballistics tables for artillery. Calculation of one trajectory took a skilled person about 20 hr. and about 3000 trajectories were needed for a single type of gun.
IV. Postwar developments
A. John von Neumann describes proposes stored program computers using binary number system (ENIAC used decimal system). Points out that instructions and data could be encoded in the same language and could be coherently intermixed in the program the machine was following and in the information the machine could store. A project was begun to make EDVAC (Electronic Discrete Variable Electronic Computer)
B. Eckert and Mauchly leave Moore School and go into business for themselves. Eventually go bankrupt and their company is bought by Remington-Rand which produces their UNIVAC (Universal Automatic Computer).
1. First one delivered to US Census Dept in 1951
2. Spectacular prediction of Eisenhower election in 1952. Computing in a different building with a fake console fitted out in the studio to impress viewers, UNIVAC predicts landslide for Eisenhower for CBS. Disbelieved and reprogrammed to give a more conservative result, but at the end of the night its initial prediction is shown to be almost right on target.
C. Jay Forrester (1918- ) MIT and An Wang (1920-1990) Harvard develop magnetic core memories for computers.
