An eon or so ago, I posted a puzzler that asked what business Samuel Clemens (Mark Twain) got into that bankrupted him. Half the people responding (that would be two people) got the right answer: a typesetting machine. More specifically, his investments in the Paige Compositor (more here) did him in financially.
This week's puzzler delves into the history of digital computing, specifically into the history of a computer's “main memory” (today known as RAM, for Randomly Accessible Memory). In the early days of digital computers, main memory was one of the biggest engineering challenges. There were no integrated circuits at all, much less the massive RAM chips we have today. Magnetic core memory was the mainstay of main memory for a couple of decades – but before core memory, there were other technologies used. For example, magnetic drum memory was used into the mid-1960s on some computers. Some of these earlier technologies seem quite exotic and bizarre by today's standards, far more complicated and less capable than seem normal today.
One of these technologies was still in use (albeit not commonly) in the U.S. Navy when I went through computer technician school in the early 1970s. Even then it seemed like a museum refugee! But I was trained to repair and maintain this main memory, though thankfully I never saw one once I left school. Today's puzzler is this: what chemical element did a once-popular main memory technology depend on?
Mercury - used as delay lines. Tubes were filled with mercury - you tap on one end, and then a fraction of a second later, the other end taps. We could also have used the Grand Canyon as an echo delay line. Add 3 + 4, shout out "7", do the next bit of math, wait for the 7 to come back for the next step.
ReplyDelete(Old-timer and sometimes student oh computer history.)
The other old memory technology that's still around even to this very day (though in another form) is the CRT - cathode-ray tube - nowadays being itself obsoleted by LCDs and their assorted ilk.
I did work once a long time ago on a machine that used drum memory - a large drum, coating on the outside, read/write heads out there too, all sealed up. One problem was that occasionally, you had to wait for the result to come back around under the read heads.