34

Over on the twobithistory Twitter channel, I came across this post from Gothamist about a computer from 1913 under Grand Central that was purportedly developed by Westinghouse.

According to the tour guide who was giving the talk,

Westinghouse in 1913, before the first World War, built the first ever, ever, ever electronic computers that could actually compute that train’s exact location.

Does anyone have more information about this mystery computer? I can't seem to find any other wiki or anything about it. I'm a bit incredulous about how this would have worked, given it seems electric switches weren't really invented until about 10 years later. I suppose it could have been an "analog" computer of some sort, maybe more like an electric calculator, but even the earliest of those weren't developed till much later as far as I can tell.

In any case, this doesn't seem to appear in any other history of early computers.

user3840170
  • 23,072
  • 4
  • 91
  • 150
dashnick
  • 1,258
  • 12
  • 17
  • 3
    IDK, but I remember seeing a list of computers that had been proclaimed "the first" by one person or another, and there were more than thirty of them on it---30+ different opinions about what the word "computer" really means. – Solomon Slow Aug 22 '18 at 20:46
  • 1
    I wonder if they had their dates mixed up. Westinghouse was an early leader in the development of analog computers, thanks largely to the work of Edwin L Harder, but as he only joined the company in 1926, and no other Westinghouse employees are regularly mentioned as having made major contributions to the field, it seems unlikely they built such a system that early. – Jules Aug 22 '18 at 20:48
  • 2
    "Actually computing a train's exact location" could be achieved by means of electromechanical relay-based signalling control. The train has passed the sensor at A but has not yet reached sensor B, therefore it is between A and B. – Leo B. Aug 22 '18 at 22:19
  • That's not really a computer though, agreed? – dashnick Aug 22 '18 at 23:06
  • 8
    To some journalists, anything that's sufficiently complex and at least marginally involves electric signals, "must be a computer". – tofro Aug 23 '18 at 09:26
  • @dashnick It's not the modern use of the term (and certainly wouldn't be a programmable computer – which tends to be the more common requirement for "serious" contenders for "first computer"), but the original meaning of computer was "one who computes" (see Human Computer) so early, purely mechanical machines "that performed calculations" could easily be referred to as "computers", even without "journalistic enthusiasm". – TripeHound Aug 23 '18 at 10:30
  • @TripeHound Except, that this wasn't computing anything. Just signaling. And I wouldn't even blame journalists here - implementations like that are just as far from todays vocabulary and understandig as some signaling in roman times are. The world - and it's vocabular - has turned way more than once since then. – Raffzahn Aug 23 '18 at 12:59
  • @Raffzahn Point taken, although it was more a general comment about the term "computer" and how it's changed over the years. – TripeHound Aug 23 '18 at 13:02
  • The earliest computers were actually ... wait for it ... people. That term was used for those (usually women) that worked out things like range calculations for artillery. – paxdiablo Jul 13 '22 at 06:13

1 Answers1

44

Does anyone have more information about this mystery computer?

Sorry to disappoint you, it's not a computer and not even a computer-like fixed-function machine. It's simply a system of cables, switches, relays and Morse writers. There was no automated signaling. It was just a bunch of switches mounted at pillars along a track, cabled in parallel. Each switch could be pulled by said cable. When a train stopped, the engineer grabbed the cable and pulled. The first pull activated the bell and powered the Morse writer. Further pulls could then be used to send over the position manually. A little clockwork was triggered with every signal; when it ran out, the Morse writer was stopped again. It was built with a mercury switch timer.

Simple but effective by any time's standards - and a great achievement back then.

given it seems electric switches weren't really invented until about 10 years later.

Reliable relays have been around since the 1850s. In theory, they could have been used to make a computer even back then - just at hard to imagine cost :))

I suppose it could have been an "analog" computer of some sort, maybe more like an electric calculator, but even the earliest of those weren't developed till much later as far as I can tell.

There were many mechanical and electro-mechanical special-purpose machinery around at that time.

In fact, 1913 and the Grand Central Station are keywords to this, with the Electric Interlocking Machine installed there. The whole Grand Central system between Grand Central station and Mott Haven (up in the Bronx) was all electric for control and handling of all switches and trains with a block based handling. Much like today. For back then, it was a masterpiece of unimaginable complexity - the whole system with several hundred sections was controlled by just 5 men in a single control room. Including a wall sized display of all sections indicating each switch and its position.

It wasn't the first interlocking system, not even the first electric, but by far the biggest of its time. While for example the Berlin subway system was several times the size, it consists of dozens of independent operating small systems, many of them still mechanical.

While the interlocking system was able to tell an assumed position down to about 1-3 miles, it was independent from above mentioned, way simpler system, and could not narrow the location down. Also it reported to the central switching room, not down to the works.

(Information mainly from memory and an article in Electric Railways Journal of 1914)

Raffzahn
  • 222,541
  • 22
  • 631
  • 918
  • 2
    No disappointment, I didn't really believe it to begin with, just wanted to learn more about it. Amazing info, thanks! – dashnick Aug 23 '18 at 01:21
  • Some of New York's subway control is from the 1930s. See https://www.nytimes.com/interactive/2017/12/22/nyregion/what-would-it-take-to-fix-new-yorks-subway.html. – Peter - Reinstate Monica Aug 23 '18 at 08:55
  • 1
    In contrast to dashnick’s comment above: Complete disappointment. And also, awesome answer. – Jason C Aug 23 '18 at 12:57
  • People forget just how high-tech and labor intensive railroading was. Up until the 1920s railroads employed somewhere around 1/10th of the non-farm labor in the US. Anything mechanical or electrical that would help them increase the number of trains they could run and reduce their labor force was pursued. – Jon Custer Aug 23 '18 at 13:24
  • I am confused about the setup, especially the cable and the meaning of parallel. You say "Each switch could be pulled by said cable.", but I am not clear on precisely what that means physically? Is there a single cable connected to all the switches? When you pull it, do all the switches move? And what does that do, exactly? – Maury Markowitz Aug 23 '18 at 13:24
  • @MauryMarkowitz Much like on a 1930s trolley. Here one cable (or rope if you like) did span several pilars. Fixed on one side, it did run thru eyes on pilars (or the tunnel wall) and fixed to the switch at the other end. When pulled, it straightened and moved the switch into closie position. On the electric side, all switches along one rail track where conected in parallel (thats an or clause to IT nerds :)). No mater which switch was pulled, curent was running, seting mercury relais to work. A mark was drawn on the morse writer. Release left a gap and release long enough stoped all. – Raffzahn Aug 23 '18 at 14:07
  • @JonCuster To be fair it's good to keep in mind that in 1920 Railroad covered next to all transportation beyone 2-3 miles. All what's today covered from trucks to aeroplanes. And in 2016, still some 8% of the US work force (farming included - but that's even more optimized today) was in transportation. So while railroads are maybe highly optimized, trucks make it up in inefficiency :) – Raffzahn Aug 23 '18 at 14:44
  • 1
    Oh I see @Raffzahn, I read it to mean the entire line was connected together by a single cable pulling all of the switches down the line. – Maury Markowitz Aug 23 '18 at 15:08
  • as an engineer, the description given for the "morse writer" sounds like an automated function triggered by an external event. sounds as though the function is carried out without human power/intervention until completion (clockwork + timer), and then resets for eventing/signaling again and again? or.. have i misunderstood? – Shaun Wilson Aug 23 '18 at 15:26
  • @ShaunWilson Not sure what you want to tell. – Raffzahn Aug 23 '18 at 17:24
  • 2
    it was a stretch, i thought this might qualify as an early data network/bus for reporting train/car positions (where cars would pull on said cable, or otherwise close the loop to signal the morse writer) but after re-reading i can see it always required a human to yank on a rope (?).. and since humans are the recipients of the morse (?) that really just makes it a clever communications device (not a network of machines as i'd hoped ;) – Shaun Wilson Aug 23 '18 at 20:49
  • @ShaunWilson It could have been more. After all, we should never underestimate the mind of our fathers. But any more comlex solution would have been overkill. For example the most simple would be one wire per section - except, isolation of wired was extreme hich tech back then and ahvign alot would be expensive. Someone coudl have overcome this by using alternatingcurent and a cleverset of diodes. Doable, but even more expensive in 1913. Simply telling the conductor to just pull the rope as often as the number on the wall tells: 3 times, pause and 7 times for section 37. Simplicity always wins – Raffzahn Aug 23 '18 at 21:06