resume

The first computer I programmed (in 1964) used vacuum tubes, the registers were stored in a Williams tube (you could actually see them on the end of the machine) and the primary memory was a magnetic drum. The machine was obsolete even then, and I was still at school.

When I got my first paying programming job, with IBM in 1967, the IBM 1130 computer I programmed had integrated circuits (some with as many as four transistors on a chip!) and a primary memory made of magnetic cores, 8 k of it (16k on the model I worked on). The removable magnetic disk was 40cm in diameter, weighed about 3kg and held 1024k bytes. I wrote a real time operating system for this machine, in assembler, and learnt the hard way what can go wrong when interrupt level code accesses the same data structures as main line code. I had to get a "Mask Interrupt" instruction added to the machine to fix the resulting mess!!

After studying mathematics at Cambridge and artificial intelligence at Edinburgh, I worked for a number of large companies culminating in 11 years at Reuters, the financial information services company. I was involved in systems design, wrote another operating system (REX), this time for the Digital PDP-11, did studies on programmer productivity, and introduced the use of high level languages for real time systems. Among the high points was a trip to Intel in Portland, Oregon in 1979 to evaluate the then new 8086 chip. "Defective architecture, not as good as the Motorola alternative", I wrote in the report, a conclusion I stand by today. Low point was my decision to reject the then completely new Unix as a potential operating system for use in Reuters. The decision was based on (a) a misunderstanding about the real time capabilities of the Unix system (b) my desire to write my own operating system.

I left Reuters in 1986 and set up my own company to market Minerva, a program I wrote under Windows (version 1.0 !) to provide real time display of financial information on wall displays. It was one of the first commercially available programs written using this brand new operating system. At one point, half the large financial institutions in London were using the system. The system, now rewritten in C++ to run under Windows NT, is still running in a number of banks.

I came to Turkey in 1991, working as a consultant and also teaching at Marmara University, Yeditepe University, and finally Bilgi University.