Ray Remembers When

Kludge Kaper Papers

Next time I'll tell you one from "How to Program a Kludge" which describes the One's Compliment arithmetic, with no designed in suppression of negative zeros, so the programmer had to check for it after every arithmetic computation and correct it if it occurred. Any wonder why a Fortran compiler for this 2 microsecond memory cycle computer ran about 60% as fast as the Fortran compiler for the 8 microsecond SDS 920 next to it at the NASA installation I worked at.


When thinking about the Kludge Kaper Papers, I always placed them as being from the early ‘60s, but remembered also that we talked about them at BNL later, after 1963, because Bob Spinrad told me that he was sure that Max Palevesky and others at SDS had written them. So, I forced myself to push my chronometer up a little. Then, when I saw that they were from the first half of 1962 I began to remember some of the specific things that I could elate to re: the machine I was working on, namely, the ASI 210. For a couple of examples: They used a Teletype BRPE paper tape punch, fondly referred to as the "Burpy" punch. The punch magnet drivers were shaky and failed frequently, and the only way to start to find the exact nature of the problem was to check the 3AG-type fuses they had used, one per punch magnet. The fuses were located by the (KKK-trained) ASI design engineers, on the steel frame between two rows of PC boards, with the separation being about 1 inch and the depth from the outside edge of the PC cards to the "push-n-turn" fuse holder being about five inches. One needed needle-nosed fingers to get at them, which was impossible without removing all of the PC boards from one of the two rows. Of course, the KKK design engineers at ASI had designed their own proprietary edge connectors, intended to be a rip-off of those designed and made by DEC. Except that the ASI connectors were so designed that the PC card could only tolerate a maximum of about 25 insertions and removals before the "grippers" of the edge connectors would destroy the etched pins on the PC card! (ASI marketing actually touted those connectors as a feature, which demonstrated the company's commitment to reliability. "The circuitry of the computer is so reliable that these in-house designed edge connectors are made so that they destroy a PC card after about 25 insertions and removals." ) And that is only one example<:))


It was so funny at the time because I had just started with Advanced Scientific Instruments, a bomb of a spin-off from Control Data. In fact, the ASI 210 was built into the same basic cabinetry as the CDC 160A. But it was a Kludge, through and through, and lived up to my expectations by surviving just a year and a half, and delivering a total of four machines, unfortunately two of which were mine to maintain, located 100 miles apart and each requiring a couple hours each day to make them run long enough to run a program. With software to match the hardware design, I had so much fun that I took the first job offered to me, which was my job with the Computer Systems Group at BNL.