User talk:Guy Harris/Archives/2022/02

Thank you for correcting my error!

Thank you for (politely) correcting my technical error in Teletype Model 33, and providing solid references for the correct info. It's been a long time since I've been hands-on with an operational ASR33, and I must have gotten its half-duplex behavior confused with another of the many terminals I've used over the years (perhaps the IBM 2741?). I thought that the topic was well worth bringing up, and was WP:BOLD in going out on a limb and writing from memory, with the idea of backing it up with refs in the near future.

You added some nice references that were not previously in the article, and which will be useful in the future. Whenever I see that you have contributed to an article, it greatly increases my confidence in its credibility and completeness. Cheers! Reify-tech (talk) 16:59, 4 February 2022 (UTC)

@Reify-tech: I was also thinking about the 2741 (where there really was a mechanical connection between the keyboard and print mechanism - I think the IBM 1050 might have had an electronic connection, but ran half-duplex only), and then about Multics, which was designed around half-duplex terminals. It was also more full-ASCII-oriented, so I don't think upper-case-only terminals were used much, if at all; they did use Teletypes Model 38 and maybe 39, but I don't remember whether the communication controllers on the host did the echoing or the terminals were run in half-duplex mode.

But researching the Model 33 also made me curious whether you might have been thinking of early Model 33s, and I looked in an older manual at Bitsavers, which seemed to speak of the Model 33 always doing local echo; I noted that on Talk:Teletype Model 33. I'm curious whether the Model 33 could always be wired for remote echo, or if Teletype changed the wiring so that it could be wired for local or remote echo, for the benefit of computer companies using the Model 33 as a terminal. Guy Harris (talk) 18:53, 4 February 2022 (UTC)

@Reify-tech: P.S. Sorry about the typo, and thanks for fixing it! Guy Harris (talk) 18:55, 4 February 2022 (UTC)

There was a bewildering array of options for the Model 33, and I remember installing the paper tape reader onto a new Teletype I had just uncrated. I was quite impressed with the mechanical analogs/predecessors to the same concepts and methods I was familiar with from the world of electronics. Conceptual equivalents of ground reference, power bus, clock timing, parallel data transmission, configuration jumpers, diode matrices, read-only memory, digital-to-analog converters, and signal connectors were all implemented purely electromechanically. Just watching the mechanical sensor pins probing the paper tape as it was being read was fascinating; if the speed had been much faster, there would have been just an undifferentiated blur.

For another glimpse of a mechanical computing system, look at the Harvard Mark I article, which I have worked on. Since it was an electromechanical computer, its main power bus and system clock was a long, sturdy shaft driven by a 5-horsepower electric motor. I actually was able to attend a ceremonial partial reactivation of the machine at Harvard in 2014, and may peripherally appear as a reflection in the crowd during a video commemorating the occasion, which was posted on YouTube (https://www.youtube.com/watch?v=bN7AdQmd8So). I actually am the voice and distorted reflection visible in another brief vignette made on the same occasion (https://www.youtube.com/watch?v=cmQ96t_YZk4).

Apparently, the Mark I was just recently relocated from the Harvard Science Center to the new Harvard John A. Paulson School of Engineering and Applied Sciences complex in Allston, so I will have to update several Wikipedia articles accordingly (https://www.youtube.com/watch?v=uYhg-CX05Rw). Reify-tech (talk) 19:56, 4 February 2022 (UTC)