000103

When Deco Electronics dissolved its California Division I joined Tasker Instruments, in Van Nuys. A man whom I shall never forget took me under his wing: Dr. Wendell Sander. He is the one who launched me into transistors. It was like jumping in the void -- vacuum tubes were all I knew. But that did not discourage Dr. Sander because he was impressed with my college thesis project, which had consisted of transforming a television set into what was in fact a crude electronic imaging system. I took to transistors like a fish to water. Soon I had my own (transistorized) analog computer, and was inventing my way into character generators -- methods of generating text and pictures for air traffic control and other displays. I was listed as one of the inventors in the patent that resulted. I needed the equivalent of a digital computer for computing the value of the resistors used in the "stroke writer" character generator, but they did not yet exist. So I devised a paper-and-pencil equivalent.

The electronic imaging system was actually called an "arbitrary function generator." It used the innards of a TV set with the deflection yoke rotated ninety degrees to obtain a vertical scan. A semi-transparent sheet of paper with a plot of the desired function (i.e., a waveform) was placed on the screen. The modulated light leaving the screen and sheet combination was picked up by a photomultiplier tube and specially designed circuits reconstructed the waveform so that it could be viewed on a regular CRT and used as an input to an analog computer.

The paper-and-pencil computer equivalent involved a human component, a Frieden calculating machine and a special form. This form had three kinds of locations on each line: locations for data recording (memory), locations for action to be taken (instructions) and locations for results (accumulator/memory.) I filled the top locations with the "inputs" and then handed the form to Sharon Stern, a technician who was good with the Frieden mechanical calculator. The interactions in the resistor-diode function generation network whose resistor values were computed by this method were so complex to analyze that it took several hours to compute the resistor values for even one letter of the alphabet.