Early Devices

I would be a bit rash to adopt the attitude that the new way is always better; however, some examples presented appear to have the advantage over methods of the past. If we think this is the case, we’ll flag it.

I addressed the subject of what “B.C.” means in a previous page. There I pointed out that it really depends on what one means by a “computer.” My personal view tends to draw the line between devices that use written programmed instructions versus those that do not. The former are “Computers” in the sense that their actions are controlled by writing and embedding the instructions in the “machine,” rather than building them into its works. I would place a machine that has been built with an installed program that cannot be changed on the “computer” side of the line of demarcation. Possibly some others would like to discuss this line?

Perhaps a simple example or two of machines that perform a real function would help those not familiar with the distinctions that I am attempting to draw.

For example, I wish to have a machine that will turn on a switch at a variable time after a defined act has occurred. There are a number of ways that men have done this in the past:

A. Have an observer note when the defined event occurs; start a stop watch; observe when the delay time has expired and then the observer manually closes the switch. This “machine” is variable in that the event and the delay can be changed by giving the observer new instructions. It is as reliable and as precise as the particular observer and the stopwatch permit;

B. However, if the event, or the desired switch action occurs in a small rocket some distant from the surface of the earth, some other means must be employed. If the event is the launching of the rocket and the switch is in the rocket, then a simple mechanical rotating machine can be envisioned that can be started at the instant of launch. If the rotating part is a cam, then when the cam has rotated a defined number of degrees, it can close a switch. We can “program” the variable time by changing the gearing of the cam, or possibly changing the voltage driving the motor. The reliability and precision of this “mechanical” machine are controlled by the selection of parts of the desired quality;

C. From this point we can depart from mechanics to electronics. A simple transistor will “turn on” when a high enough voltage is applied to its control point terminal. At any voltage below the “turn on” point, it does not conduct electricity. If we design a simple circuit using the transistor as a switch and apply a variable voltage to its control point, we can cause the voltage to reach the “turn-on” threshold at the precise time desired simply by selecting a resistor and condenser of appropriate values in the variable voltage circuit. Again, the reliability and precision of this “electronic” machine are controlled by the selection of parts of the desired quality;

D. On the other hand, we could choose a small computer that has the capability to control several functions using a variety of different input signals ranging from rocket launch to receipt by radio of a variety of commands. This computer could have all these functions embedded in the machine at the time of manufacturing, or it could have the capability to change the commands at anytime prior to launching the rocket, or it could receive new commands after launch by radio signals. In the earlier days in space, when such items became available, these machines would have been used cautiously because of their lack of demonstrated reliability. Use would have required triple redundancy and a way to determine which of the computers was in error.

Which of the methods used for the rocket may have resulted from when the launch occurs. In the early days of the “space race,” option “D” was not a reality; however options “B” and “C” were definite possibilities. Even after option “D,” the general purpose computer was available, it might not have been used for very simple functions, since it is much more sophisticated, costly and heavy than the other available solutions. Only where other requirements, such as numerous other functions needed to be controlled, or the need to change command parameters while in flight, would the small computer have been selected. Even today, simple computers require significant testing for the space environment and reliability. Redundancy is generally required; how much redundancy depends on the vitality of the functions they perform.

Devices “C” and “D” are on opposite sides of the line that I draw between “BC” and “AC,” after computers.

I hope that these examples clarify the dividing line.