Tuesday, May 19, 2020

Debugging and emigration

What was wrong with the SM-1800?

That was a Soviet clone of the PDP-11 (Yes, I remember the PDP-11: we used it and a smaller LSI-11 for data acquisition.). It was routing trains and cargo, and was crashing at night.

Read the story.

I'm a little puzzled--perhaps they cloned a version of the PDP-11 that didn't use parity memory.

5 comments:

Douglas2 said...

This tale has been pasted on many fora, and it encountered some skepticism on Reddit:
https://www.reddit.com/r/programming/comments/dquu3/debugging_behind_the_iron_curtain/

And while many of the Soviet SM series of minicomputers were PDP-11 clones, my information is that the SM-1800 was not.

james said...

As you can guess from the second link, I was puzzled at the lack of parity memory, and went to look up the matter. It turned out that my hazy memory was correct, and that parity bit memory was an option--you could buy memory cards without it.

Even so, the radiation levels to cause that kind of damage are quite high--not the sort of thing you're apt to get with cattle; they'd die first. The train cars themselves might have gotten contaminated, though.

Douglas2 said...

There's a comment on the reddit thread mentioning dust on the railcars being stirred up by unloading and then blown-through the cooling system for the minicomputer - it's a conceivable but implausible means of (maybe) getting a strong enough field close enough to critical parts. Not as outlandish as some other tales of debugging in folklore collections: https://beza1e1.tuxen.de/lore/, so perhaps possible?

Other tales of untraceable provenance that tell of Chernobyl radioactive cattle:
• have it being slaughtered within the region, with the less-radioactive meat being mixed in small amounts with other meats to make sausage which was directed to be sold throughout the union "except Moscow", and
• the more radioactive meat bing put in cold storage until the radioactivity decaued enough for use.
• The story goes on that the region ran out of space in commercial freezers, and started using refrigerated boxcars, which then spent the next many years being moved around the Soviet union as each depot rejected delivery, and finally the train went back to the exclusion zone where th emeat was landfilled.

So there are elements of another folklore tale, the difference being that the cattle are live cattle in this one. IT also parallels one I've heard debugging by train-schedule-analysis of new trains in Singapore (I think) having random "fail safe" safety stops when in service, never repeatable at time of day or location or train carriage in use, with programmers working weeks to figure out the problem, and finally discovering that one of the new trains was had excessive RFI emissions and would occasionally cause a faults on another of the new trains when the two were passing on adjacent tracks

james said...

Sometimes the stories are true:

"The real culprit, the TGV (a high-speed train), was discovered by accident a few weeks later during a discussion with a railway engineer: leakage currents on the French rail track flowed through the LEP vacuum chamber with the return path via the Versoix river back to Cornavin. "


I can believe that one tranche of cattle got contaminated, but not more. But if you told me somebody grabbed some train cars from downwind of Chernobyl without testing or cleaning them...

james said...

But yes, I've no way to track this down to see if it is true.