There was a private request for another early 60s computer story, and the UNIVAC tale had a sequel that might entertain some funsters. We were not allowed to list classified data on the Flexowriters because each character's binary code energized a distinct set of solenoids, broadcasting a signal that, in principal, could be decoded at a considerable distance. I seriously doubt (but dearly hope) the KGB expended the resources to intercept my Simpson's rule printout. If they ever figured out how it worked, the mathematics contained a completely fallacious model of our sensors, and the fancy programming gave the monstrous misimpression that we knew what we were doing. In fact, the Systems Analysts flatly rejected my computation, and very rudely told me to go back and try again. But I was sure of my results, and even surer that the analysts were bozos. These were the guys who spent over a year coming up with a way to intersect three ellipsoids. One of them once accused me of "ruining the randomness" by scaling uniform Monte Carlo variables instead of retrying until I got one in the desired range. He later demanded to know why the sum of two logarithms wasn't the log of the sum. I finally got some high-level director to force the bozos to show me the math behind their disputed integral. It turned out they were using a decaying Gaussian to model a distribution whose actual density was zero outside an annulus with an expanding hole whose center was the mode of their Gaussian. --rwg