Jon Elson's Computer History

LINCs, PDP-8s and IBM 360s

I started at Washington University as a freshman in the fall of 1969. I quickly hooked up with a nerdy bunch of engineers, mostly EE's and computer science majors. Washington University had a lot of computer systems development going on at that time, a lot of LINCs (Laboratory Instrument Computer, a 12-bit machine with 2K words of core memory) were in use there, there was Biomedical Computer Lab at the Med School, and a lot of medical imaging projects going on on primitive gear attached to the LINC's.

My freshman advisor was William Papian, who was the grad student who helped Jay Forrester develop core memory!

I was fascinated with all this stuff. For regular class work, we used the university's IBM 360/50, with achingly slow turnaround of batch jobs submitted on punch cards. A really painful way to learn programming, but it did teach you to bench check your program before submitting it.

I got into an advanced comp sci class on image processing, and we got to use the only Spear MicroLINC on campus, just a faster version of the LINC. The MicroLINC used early Motorola ECL, the classic LINC was a discrete transistor machine. It was really cool to manipulate a 3D wireframe image on the screen, but the complexity of the image was SEVERELY limited by the speed of the LINC. A cube with one notched corner, so I think that is 10 vertices and 13 vectors bogged it down to about one frame per second.

LINC console, tape drives, etc.
LINC main cabinet, with door open, showing half of the backplane wiring

We had a cast-off DEC PDP-5 computer (basically an early version of the PDP-8, all discrete transistors) in a pretty much defunct lab called the Hybrid Computer lab, which also had several PACE analog computers and a really decrepit Heathkit analog computer. The PACE computers looked amazingly cool with brightly colored inserts in the panel to identify functions, but were pretty much obsolete by 1970 or so. Leonard Berger convinced the university to accept a donated Bendix G-15 from NASA, they just had to pay the cost of shipping, and hook up power for it. The G-15 was a serial-arithmetic, drum memory, vacuum tube and semiconductor diode machine of really low performance. We all thought it was a true boat anchor, even though there was an early Algol compiler for it, it was so SLOW that it might take DAYS to run a simple program. It had almost as much memory as the PDP-5. Nobody was able to get it to do much, and I eventually removed the cover on the memory drum and found the drum was badly scored. One might still get it to do something, but any major software wouldn't know to avoid the bad tracks on the drum. The only I/O on the thing was a typewriter and paper tape.

I worked part time at Artronix, which was building a 12-bit computer mostly used for radiation treatment planning. This was a descendent of the LINC, mostly it just had expanded memory and a memory remapping unit called the origin register. I mostly wrote diagnostics for it. They built everything there except the core memory module. They etched the PC board themselves.

Then I went to University of Missouri at Rolla, another 360/50 but that was mostly used as an RJE terminal to the 370/168 at the Columbia campus. Took a few programming courses there, including 360 assembly language and wrote what was essentially a simple language interpreter as a group project.

Rolla also had a bunch of Data General Nova computers (16 bit minicomputer) that were available for just fooling around. I wrote a bunch of programs to learn about stuff I was interested in, like linear feedback shift registers - also called maximal length pseudo-random sequence generators. I met some really cool guys there, too, such as Walter Lewis and Doug Wood.