Recently, there was much discussion on my social media pages about ancient computers and teaching assembly language with regards to these older architectures.

I have long advocated that even in the days of virtual machines and Java that computer science students and computer engineering students should learn at least one assembly language early in their programming education. This, in my mind, is necessary to avoid mistakes in program design and to be able to discover errors in compilation that might occur due to the compiler or optimizations requested.

Perhaps some of you are not aware of the magnificent work of Bob Supnik, a vice president and engineer who I had the privilege of knowing and working with for a number of years at DEC.

Even before he left DEC, Bob started working on an emulator now called SimH [1], which runs on a variety of different operating systems and emulates an even larger number of hardware architectures. Not just DEC architectures … many different architectures.

Using this emulator, you can run old binaries of operating systems and programs on your laptop, and, armed with hardware manuals and old literature off the net, you too can program these old machines in assembly languages and run old operating systems, compilers, etc., to your heart's content.

But wait! There's more!

Those of you who know me know how much I loved two architectures: The PDP-11 for its elegance and the PDP-8 for its simplicity (and because it was the second computer I ever programmed, and the first assembly language I ever learned). However the usefulness of learning machine/assembly language is amplified by learning both the PDP-8 and the PDP-11 architectures, to see the difference between the two.

The PDP-8 was enormously (for its time) successful, for it could fit on a tabletop and was fast enough to process data, mostly from engineering/scientific types of problems. The PDP-8 had only 4,096 12-bit words and could not even subtract. To subtract you had to take the two's complement of the subtrahend and add it to the minuend (I have not even thought of those terms in about 20 years), and then check for overflow.

The PDP-11, the predecessor to the VAX line of computers, was probably one of DEC's most profitable and longest lasting architectures, with its refrigerator sized racks of circuit boards and (often) core memory.

Recently I learned about the PiDP-11 [2] and PiDP-8 [3] projects, the loving work of Oscar Vermeulen, where Oscar recreated the front panels of a PDP-11/70 and a PDP-8, complete with switches and lights, and made them work with a Raspberry Pi running SimH.

You do not need these front panels to explore these architectures, but it makes the exploration that much more interesting and fun … to see the lights blink, to read the registers and memory through them, and to actually change values in the memory through the front panel.

The project includes a directory of different operating systems to run on the PiDP-11 (RT-11, RSX-11m, RSTS/E v7, and a multitude of different Unix systems). The PiDP-8 has operating systems and compilers too, and both have an assortment of games available, although you should not be expecting high performance graphics.

But wait! If you want to really get "down and dirty," there is the Gigatron [4]. The brainchild of Marcel van Kervinck and Walter Belgers, it is a complete CPU and system built out of simple logic gates, the types used before the days of single-chip microprocessors.

This allows you to see how the machine language ones and zeros actually control the CPU, and how few transistors (also known as "gates") can be used to make a CPU.

This is not a machine that can only be programmed in machine/assembly language. You can also run BASIC on it (although you are somewhat limited to the size of the program and variables). There are also many videos, tutorials, and articles on the Gigatron talking about how it was designed and some of the design trade-offs taken.

The Gigatron can also play games, although (again) do not look for 3D graphics – it is estimated at the power of an Intel 8088 computer.

I often tell audiences that people do not buy hardware, and they do not buy software; they buy solutions. People do not glue hardware and boxes of software to their walls with candles on either side like a shrine.

In this case, I will make an exception. I can foresee a small portion of my wall where a PDP-8 and a PDP-11/70, as well as a Gigatron may happily whirl away as a shrine to the early days of computer science, but at a fraction of the electrical power (and cooling) that the original machines used.

Thank you gentlemen, for all your work in recreating these "old" machines.