On Saturday, September 19, 2020, 11:53:08 AM PDT, grarpamp <[email protected]> wrote:
https://bsdimp.blogspot.com/2019/10/video-footage-of-first-pdp-7-to-run-unix.html When I began MIT in 1976, I very frequently visited the MIT electronic club, MITERS. In a room that no longer exists, as I recall 20B-114. https://en.wikipedia.org/wiki/Building_20 They had a PDP-7, it might very well be the one in the video? They also had a PDP-1, originally given to MIT in gratitude by DEC, for MIT's assistance in designing said computer. These two old DEC computers were traded, with DEC, for 40 pallets full of what DEC probably thought of as "junk" , but we thought of as "junque". Probably an 11/23, an 11/34, and maybe even parts of an 11/70. One condition: MITERS had to take IT ALL!! Apparently, these machines were being collected for the Computer History Museum, which a few years later (1983?) I visited in Boston. I believe the main guy at MITERS then was Jordin Kare. Sadly, I just found out through Wikipedia that he died in 2017. https://en.wikipedia.org/wiki/Jordin_Kare One curiosity of that specific PDP-7 was that somewhere on the wire-wrapped backplane, there was a very short wire, wire-wrapped, maybe 3 inches long, and leading to...nowhere. There was a note on it that stated that if that wire wasn't present, the machine simply did not work. WITH that wire, the machine...worked. No other explanation. Presumably, a slight capacitance, most likely discovered by applying an oscilloscope probe during debugging, and noticing that the machine worked only when the scope probe was applied. The technology of the PDP-7 was once explained to me as "40-nanosecond pulses chasing each other inside the machine". I think it used germanium transistors. I believe this specific machine had 8 kilo-words of maybe 18-bit magnetic-coremain memory, plus a couple of DECtape drivers, an X-Y vector display. It played a game called "Spacewar", a 2D "four spaceships around a sun" shooter. https://en.wikipedia.org/wiki/Spacewar! In the front panel, there was a knob leading to a rheostat, which adjusted the clock-speed of the CPU. The idea was that you'd turn the speed up until it began failing, and then you'd turn it back down and run with that. Jim Bell
