In later 1972 I attended a seminar given by Intel Corporation. The seminar was designed to introduce engineers to Intel's new 8008 CPU-on-a-chip.

It is important to understand that the idea of a complete 8-bit CPU-on- a-chip was an absolutely revolutionary concept at the time. Electronic engineers were firmly entrenched in the practice of designing digital circuits to implement whatever logic was needed to perform a desired function. The design of a complex piece of equipment using integrated digital circuits could involve months, even years, of work.

Much of this work consisted of painstakingly "breadboarding" the circuits in prototype form in order to test the design. This involved connecting hundreds to thousands of wires between device packages in order to confirm the design.

In designs of this complexity there is ample opportunity for error. Wires are connected to the wrong points. Schematic drawings get drafted incorrectly. Test results are lost. Changes have to be made. Revised versions often need to be constructed.

Crafting a circuit board containing a couple of hundred integrated circuit packages (and thus several thousand wiring connections), checking it out, and finalizing the documentation for production purposes could easily cost thirty to forty thousand dollars in 1972. Thus, the development costs for a product using a dozen such circuit boards could easily cost a small fortune.

Electronic equipment intended primarily for the control of machinery and systems could be designed to utilize a computer. But, at the time, even small so-called minicomputers cost thousands of dollars. They also occupied cabinets taking up many cubic feet of space.

The advantages of using a computer to control a system were that a design could be accomplished in less time, could be easily tested and changed, and would remain flexible into the future. To change a design you simply changed a "program" within the computer.

On the other hand, changing a digital circuit usually meant throwing it away and manufacturing a new one.

However, in the 1970 era, using a digital computer to control small systems simply was not practical from an economic:: viewpoint. For, while a typical control system might be designed to use a computer at a development cost of, say, fifty thousand dollars, the small computer used in such a system might cost two thousand dollars each. If you planned to manufacture ten thousand systems, you would need to invest twenty million dollars just to acquire an inventory of computers. On the other hand, while a digital circuit implementation of the same control system might cost half a million dollars to design, those same circuits might be manufacturable at a cost of just 500 dollars per system. That would mean ten thousand such systems could be produced for something on the order of five million dollars instead of twenty million dollars.

Thus, while designing with digital logic circuitry was slow, expensive, and difficult to change, when viewed over the life of a mass-produced product, it was considerably cheaper than using a digital computer at that time.

Of course, what was needed was a digital computer that was small and inexpensive. Intel's model 8008 CPU-on-a-chip was the first such device capable of handling data in 8-bit chunks to become commercially available to electronic engineers.

After studying the literature given to me by Intel personnel at the seminar on the 8008 CPU-on-a-chip, I quickly became convinced that I could use the 8008 to replace a great deal of the logic chips I was using in the design of a product underway at the firm where I was then employed as an electronic engineer . There was no doubt in my mind that chips of the revolutionary type represented by the pioneering 8008 would be the wave of the future.

Indeed I was completely enthralled by the exciting potential of digital computing at the time. This was evidenced by the fact that I had purchased a Digital Equipment Corporation PDP-8 computer for my own experimental use at home. This was a monstrous machine, residing in a six foot high metal cabinet (somewhat larger than a typical household refrigerator). I had in esconced in a room of the house I was renting, much to the amazement of friends and neighbors. It was coupled to an old clackety model 33 Teletype machine, equipped with a paper tape reader. This was used for input/output. The PDP-8 was originally sold by DEC for about twenty thousand dollars. However, I obtained my refurbished unit for about a tenth that cost from DEC during a sale it was having to dispose of obsolete models. It was a pretty hefty investment for me at the time. Not too many individuals had purchased these machines for personal use, as evidenced by the DEC representative who handled the sale. He intimated he had never heard of a PDP-8 being installed in a persons home.

One of the first programs that I developed on the PDP-8 was a cross-assembler for the Intel 8008 CPU. That was a program that would assist me in the assembly of code for the Intel 8008 device. Soon I was busy experimenting with various kinds of routines for the 8008. Much of this early experimentation was designed to exploit the capabilities of the 8008 for a system I was designing for my employer at the time. I had no doubt that the use of that CPU-on-a-chip could save my employer a great deal of product development time as well as reduce final product costs.

I was also young and naive enough at the time to believe that my exhortations to use the 8008 in products I was working on would be quickly adopted. Management at the company I was employed by considered the firm to be on the leading edge of technology and appeared to want to stay that way. It was using programmable logic arrays at the time (they were in their infancy) and was exploring the use of large scale integration technology.

Alas, my enthusiasm for the 8008 was not appreciated by management. There were general concerns about such matters as second sourcing the device. And, as forward as the company was technically, it could not seem to grasp the significance of the device at that time. (They did a few years later when competitors attacked their product line using microcomputer based equipment!)

I was severely disappointed by the lack of enthusiasm on the part of my employer to support my interest in using the 8008. When it became apparent that I would not be able to work with the 8008 under the auspices of my employer, I began to look into using the device on my own.

An initial goal was to obtain a supply of the 8008 chips and the random access memory circuits designed for use with the CPU. I soon wrote to Intel and asked if they would be interested in licensing the use of my PDP-8 cross-assembler package for the 8008. I suggested a trade of the rights for about two thousand dollars worth of their chips. A representative of Intel soon countered with an offer of something like three 8008 CPUs and handful of memory ch1ps, worth considerably less than what I had suggested in my proposal. I decided to decline what I considered to be a rather stingy offer on their part. I could always market my cross-assembler myself, I figured. And, at two hundred dollars a copy, I planned to bring in considerably more than Intel was apparently willing to offer for the package.

Soon thereafter I convinced two other engineering associates to join with me in constructing a set of three prototype 8008 "personal" computers. We each chipped in about two hundred dollars to purchase the necessary 8008 CPUs and a couple kilobytes of static memory devices.