The Early Days of Computing

At the very beginning of stored programs, and thus software, computer hardware was relatively large. Nevertheless, only one person could use a computer at a time, making a computer the biggest personal computer ever.

There were several hundred years of frozen-programmed or hard-to-program mechanical analog computers, most in the 1600 to 1930s period, before the early electronic digital computers. As an example, Dr. Helmut Hoelzer designed and produced a complicated analog computer to help guide the A-4 (or V-2) rocket [Tomayko85]. However, due to the specificity of electronic analog computing, the actual guidance system could not be used for other rockets , although the principles are the same. Similarly, the analog computers used for the flight control system on the General Dynamics F-16A and B were not transferable exactly to the Lockheed F-117A, but the latter airplane s flight control system used the same analog types of circuits in a different arrangement (later, a version of Hoelzer s computer was transferred to several U.S. launch vehicles) (see Figure 8.1).

Figure 8.1: A Redstone rocket lifts off (using guidance vanes in the exhaust). Redstone Arsenal Historical Information

Later electronic digital computers could be operated via software. Most, like the Electronic Numeric Integrator and Calculator (ENIAC), were still difficult to program or could hold only one procedure at a time. While University of Pennsylvania faculty, like John Mauchly, and graduate students, like J. Presper Eckert, built the ENIAC, meetings were held between them and the famous mathematician John von Neumann to look ahead. One thing that was immediately obvious to them was that without the ability to easily change the program, electronic computers would be nearly useless. At one of these discussions, the participants realized that programs and data could be represented using the same techniques used in computers.

Von Neumann wrote, in 1946, A Draft Report on the EDVAC, which, unlike the ENIAC, used binary representation of numbers (the ENIAC was a decimal machine). The stored program concept described in this paper was derived by the group at Penn collectively.

Eckert and Mauchly left Penn to form a company that eventually built the Universal Arithmetic Calculator (UNIVAC). The EDVAC design formed the core of the Johnniac, named after von Neumann, at Princeton University s Institute for Advanced Study. Meanwhile, the idea crossed the ocean, coming to fruition first in England. Manchester University built the little-known Baby computer with stored program capability in 1948. Maurice Wilkes of Cambridge University s Mathematical Laboratory built the much more capable, and better known, Electronic Delay Storage Automatic Computer (EDSAC) in 1949.

The EDSAC s memory, a big-for-the-time 512 words of binary numbers, was acoustic . This was accomplished with glass tubes mostly filled with liquid (room temperature) mercury . One end of the glass tube had a speaker of sorts, and the other end had an amplifier . Ones were stored as high amplitude sound waves in the mercury. Zeroes were low-amplitude waves. When ones or zeroes reached the amplified end, they would be strengthened and returned to the other end (see Figure 8.2). This is called a Delay Line , since a number took some time to be read. When one number needed to be read, a counter circuit would count the numbers as they went by in a specified bank. Eckert devised a similar memory scheme during World War II.

Figure 8.2: A Mercury Delay Line memory.

Tasks  

1. What is the stored program concept? What is the important insight in this observation?

2. Find out what a Williams Tube memory is. Discuss whether it is a better technology than the Mercury Delay Line.