History, Page 2

While still at Hughes Aircraft Company about 1955, I became aware that a computer needed only one flipflop if it had the appropriate simple memory.  My description of such a machine appeared in 1958 in the article "Logically microprogrammed computers" in the IRE Professional Group on Electronic Computers, Vol. EC-7, No. 2, pp 103-109. Thus, I began to think there could be simple computers which could be afforded by individuals. This was in the days that even a small computer cost tens of thousands of dollars and more. I was frustrated though by the lack of a memories that I could use. The secret to the design of such a simple computer was to store the description of another computer in the memory and then to evaluate what the target machine would do when faced with a program. With one flipflop and the department head's statement that a flipflop added $500.00 to the cost, I was thinking of a computer for personal use for $500. This was my vague target.

In the fall of 1970, I found myself unemployed and I decided to investigate what might be done to make a computer for personal use. My criteria for the computer were low cost, educational, and able to give the user satisfaction with simple programs. The computer could be serial and slow which would reduce the cost yet create the environment that was desired. It should demonstrate as many programming concepts as was possible. Because of the small size, the native language of the unit would be the machine language. Above all, it had to be a stored program machine in the von Neumann sense. At no time, did I consider designing the computer in accordance with the principles of logically microprogrammed computers. Instead I decide to apply my talents to the design of a simple computer. To keep the costs low, switches and lights were the input and output of the machine. (Some thought was given to punched card input but it was never developed.)

The design emphasized the use of standardized parts as much as possible. The initial objective had been a parts cost of $150 per machine but the actual figure was closer to $250. (In volume production, the figure of $150 might have been met.) By the spring of 1971, the logic printed circuit board had been built and the computer was assembled. The memory was two MOS shift registers, each of 1,024 bits. The logic was implemented with standard small and medium scale integrated circuits. The microprocessor had not been announced yet nor would the first one been an improvement in the design.

Professionals were enthusiastic about its characteristics and everyone agreed that it was educational. Perhaps because the computer was so educational, the marketing emphasized schools and the majority of the units sold by Kenbak Corporation were to schools. (Perhaps it would have been better to emphasize the fun aspects and to have pitched the marketing to individuals.) There was a problem with selling to educational institutions and that was their lengthy budget cycle.