~Oswegan

I’m actually amazed that it was dismantled after the war. I mean, couldn’t it have continued to be of use to intelligence bureaus ? Or was it seen as some sort of dead-end?

It was purpose built to speed up cracking of the Lorenz cipher, and once this was no longer needed, colossus could not be repurposed to other tasks.

I wrote an essay a while ago comparing Colossus, Mark I and Zuse Z4, if your interested in early computers.

If you do visit, try and find a day when Tony Sale is there and leading the tour as his knowledge and enthusiasm is wonderful.

If you want any local info please feel free to drop me an email.

Thanks

Like the Funk Brothers, it’s good to see, at long last, these erstwhile chappies getting their nod. Pip pip chappies!

Dybbuk: for the real American fifth column, read up on British Security Coordination.

Mikelotus: I always understood that Turing was treated pretty well at Bletchley. It was post-WW2 Army strictures and archaic anti-homosexuality laws that got him.

In any case, at the time they had a version of the code breaking system runnin on a laptop (I think it was an older powerbook) and the Colossus, once it had warmed up, kicked the laptop’s ass.

I think it’ll be an interesting contest.

Oh wait, you mean the *other* Colossus…….

That was a pretty cool machine. It was used to decrypt German teletype codes, as opposed to the better known Enigma radio codes, once the Alliies were in the E.T.O. after D-day. They originally tried to do this with a mechanical system that spun teletype tapes around at 30 miles per hour XORing German messages with those same messages offset from the originals to varying degrees. You knew that you had a good guess when some section of result had a lot of zeros in it. Problem was that the tapes tended to stretch so you often wasted runs because the slippage screwed with the banks of photoelectric cells that were watching for a good run of zeros.

So they expanded on the idea of using vacuum tubes that they’d come up with in the course of building one of the monstrous Steam Punk engines that was working on the Naval Enigma problem. Over there they had a set of six mechanical Enigma “Bombs” (mechanical simulations of Enigma wheel settings as applied to a specific intercepted message) being cranked along through a series of related key guesses by a central shaft at several hundred RPM. Once a regular pattern was detected in their output you had to stop the whole works and read off the current set of guesses to some human who had to see if the regular pattern amounted to German Military Speak or was just statistical clustering. Problem was that mechanical relays often were too slow to keep up with the whirly gears at anything past half speed. Even then a relay might stick and produce a false positive that could waste a half-hour while you applied the brakes and then carefully rewound to the spot where the the pattern had seemed to emerge.

So they invented the idea of using vacuum tubes as memories storing the instantaneous output of the Bombs for the fraction of a second that it took for the Pondering Circuits to see if the pattern that they were looking for was there. They were, even back then, more reliable than the best relays so you saved a lot on time and brake pads due to lower number of false positives. Best of all, you could really pour on the horsepower to the central shaft as the tubes could jump to each new set of outputs way faster than the relays could.

So they went from the idea of storing readouts from physical reality in tubes to storing simulations of physical reality (the teletype tapes and their relative positions) in Colossus. Thus was the first VR born!

The Mark I had 1,500 tubes but later models got up to 2,500.

(See Battle of Wits, by Stephen Budiansky.)