My introduction to microprocessors was in 1979 when I built a "Science of Cambridge MK14" kit. This was based on the INS8060 microprocessor known affectionately as "Scamp" from SC/MP, simple cost effective microprocessor. The basic kit had 256 bytes of RAM and 512 bytes of ROM which contained a machine code monitor to communicate with a twenty key hex keyboard and a tiny 8 digit seven segment LED display. I ordered the expanded kit which had an extra 256 bytes of RAM and an I/O chip, INS8154 which has 128 bytes of RAM, two eight bit I/O ports and some control lines. My first priority was to make a tape interface so that I didn't have to enter hundreds of bytes through the keyboard each time the power was applied. The monitor had the necessary save and load routines and I found a design for the hardware published in "Practical Electronics", a magazine I had subscribed to since 1974. The first application I wrote decoded the fast code of MSF Rugby's 60 kHz atomic time signal and displayed the time on the LED display. I received the signal using a Datong UC1 converter, an Eddystone EA12 receiver and the tape interface which turned the audio into digital. My next project was a morse code trainer. This generated a random stream of morse code at any programmed speed with weighting on 'difficult' letters if required. Shortly after writing that software, I applied for a job as an assistant development engineer with Lucas Micos Ltd. in Cirencester, England. The two engineers who interviewed me were impressed with my self training and the projects I had described and told me that I had the job.
I left Lucas Micos to work for DKS Automation in Corby in early 1982. There I was fortunate to work closely with the technical director, Bill Kilkerr, the "K" in DKS, who designed the hardware for projects and Tony Skelding, a freelance software engineer. Tony delighted in teaching me the intricacies of his software. I was particularly fascinated with a multi tasking scheduler he had written in 1802 assembler for a project which controlled a silver smelting furnace. The idea of concurrent running processes crystallized when I correlated Tony's low level scheduler and the high level functions of Texas Instrument's real time "Microprocessor Pascal" for the TI9900 series of processors. Later, DKS switched to using Forth running on the 1802 rather than pure assembler. They bought a Forth development system from Microprocessor Engineering in Southampton. This was, in modern designation, a pre alpha release and Stephen Pelc, the founder of the company, came to DKS for two weeks to get the Forth nucleus running. He left at that point and it was up to me to debug the Forth primitives and extend the package with arithmetical and I/O functions. I was made redundant shortly afterwards when DKS's parent company went bankrupt. Stephen directed companies in my area which enquired about on site support from him to me because he didn't want to travel a couple of hundred miles and stop work in Southampton. All of my subsequent contracts were as a result of personal contacts and I never obtained a contract through an agency.