levin@bbn.com (Joel B Levin) (10/13/89)
I have some additional comments on the general topics of Western Union, teletypes, clocks and telegraph offices. I visited a telegraph office once, on a school field trip over thirty years ago, in a small city (about 17,000 and a university). My recollection is of the standard marble top counter with the pads and the machinery behind. One thing no one has mentioned is what I think was the standard form of a telegram in those days: the telegram was printed not on the form, but on a long strip of gummed paper. This was fed through a little dispenser (not unlike those sometimes used today for taping up boxes) which allowed the operator to simultaneously gum the strip and apply it to the form. It had a serrated tip so the operator could tear it at the right edge of the paper and start a new line below at the left. It is my recollection that telegrams were often prepared much the same way that we much later did Telexes: a paper tape was punched off line, then read through at speed (such as it was). This of course allowed the correction of errors. Telegrams had a feature (and may still for all I know) of repeating in a string all numeric words at the bottom of the telegram, including any figures in the body of the telegram as well as street addresses and postal zones in the header. Presumably this was done for reliability, since transmission or transcription errors are much less tolerable in numbers than in text. There has been some discussion of the 5-level teletypes. I played with a Telex machine for a time, though I had already had a lot of experience with paper tape on the 8-level Model 33 (as a minicomputer I/O device). As delete (then called rub-out, for good reason) was used in the ASCII devices to obliterate errors, since it was coded as all eight holes punched, LTRS, which had all five holes punched, was used on the Telex machine. This reset the device to its unshifted state, was otherwise a "no-op" character, and so was used as leader and trailer on a strip which contained a message. [There's a lot more that comes to mind on the topic of those mechanical beasts, but I'd better pass.] As an undergraduate, I too worked in a radio station which had Western Union clock service for the first couple years I was there. Our station also had the hourly signal wired through the control panel to an oscillator so it could cause a beep on the air if enabled (since we broadcast a lot of classical music, we didn't want it on all the time). It was a skill practiced by the announcers to "make beep" following any program that ended on the hour. In particular, one had to know exactly when beep would occur. The clock lost perhaps a half second an hour, so you knew beep would happen when the second hand was one and a half ticks away from the top; if you were going to make it, you could signal the controlman (we used a two person operation) who would push the button to enable beep. Then at just the right time, the red light on the clock would flash, the second hand would leap to the top of the clock and freeze there for a whole second, and the world would proceed as usual. If the WU line was down for a few hours, it became a little tricky to figure out at what second beep would come when it was restored, as the clock might be off a little by then. Then we came back one year, a bland white clock was on the wall, and beep passed into legend.