From the February 2014 issue of Futures Magazine • Subscribe!

Building a better profit-catching mouse trap

Computer technology has always advanced in stages, and at each stage of advancement, it has taken software developers some time to catch up to the leap in hardware. We now find ourselves at just such a pivotal moment, with multicore technology here in force and the cloud-based processing on the horizon.

This is a point in the evolution of trading technology where the hardware takes center stage. In recent issues, we examined how software has evolved over the years. To come to a full understanding of where we’ll go from here, it’s necessary to have the same discussion with respect to hardware. Then, we may better grasp where we’re headed and how traders will benefit from new designs, programming and computer technology.

Here, we will begin that journey with a trip down computer memory lane, examining the key leaps in both processing and operating system technology. In the next installment, we’ll directly relate future trading technology to the next stage of computer design.

The beginning

The IBM 360 operating system was the first operating system to encapsulate all aspects of a modern O/S. There were three versions released to the public. One was like prior mainframe computers in that it could only run one job at a time; this was known as PCP, or a Primary Control Program. Two other models allowed multitasking. The first was called MFT, or Multiprogramming with Fixed number of Tasks. High-end machines could use MVT, or Multiprogramming with Variable number of Tasks.

Although the MFT and MVT versions had their day, other programs ultimately won the consumer desktop battle. Many base the history of the modern PC with Tim Paterson in June 1978. Paterson was a computer science graduate from the University of Washington and took a job at Seattle Computer Products as a designer and engineer. He designed the hardware of Microsoft’s Z80 SoftCard, which had a Z80 CPU and ran the CP/M operating system on an Apple II.

Meanwhile, engineers at Intel were busy working on a chip that would be the death of the Z80 processor. They called it the 8086 (also called the iAPX 86), a 16-bit microprocessor chip released in mid-1978. This chip formed the foundation of the x86 architecture now found in almost every computer in the world. When this chip was released, Paterson went to work designing an S-100 8086 board. Software support was weak, so he developed QDOS (Quick and Dirty Operating System) in April 1980. QDOS was soon renamed as 86-DOS, and version 0.10 was complete by July 1980. By version 1.14 86-DOS had grown to 4,000 lines of assembly code. 

The 8086/8088 chip combined with the OS allowed for hardware interrupts, which jumped to code in low memory. This meant that when these events occurred, a given block of code was executed. A time interrupt would run at every fixed interval. For example, these key low-level interrupts were used in developing print spooler technology by hooking the timer interrupt and using it to create primitive multitasking applications with a technology called TSR, or Terminate and Stay Resident.

Sidekick was an example of a program that used this TSR idea well. Released in 1984 by Borland, this program could be loaded into memory by pressing down Ctrl-Alt. When it was done, it returned the user to the DOS prompt. While DOS was a single-task environment, this had the effect of mimicking multitasking. Sidekick came with a few useful applications such as a calendar, text editor, calculator and other basic utilities.

comments powered by Disqus