That UCLA differential analyzer from 1947 was the peak of that technology. Earlier ones in the US go back to 1928. Wikipedia says UCLA had three of those. Probably for the Southern California aircraft industry.
That was near the end of the line for the pure mechanical analog computers. Even the gun and missile guidance systems had electrical inputs and outputs. 1947 was late to be building that, but it was a mature technology and worked, although, as shown, a huge pain to reprogram. Mechanical fire control systems had one built-in problem to solve, so they didn't have to be set up as a parts kit like this.
+1 for the narrator using the traditional "Los-an-guh-lus" pronunciation instead of today's "Los-an-ju-less," popularized by The Brady Bunch and other 70's television programs.
I prefer "Los-an-ju-leez," as that was a common style among radio announcers. But all three are officially correct, according to an LA Times article I read years ago.
Also, if you're into this sort of thing here's the first in a seven video series on a very different mechanical computer - Albert Michelson's Harmonic Analyzer which does Fourier Analysis: https://www.youtube.com/watch?v=NAsM30MAHLg
I love the tone of amazement and excitement in the video. The images were amazing, but the tone really stuck with me.
I wonder what could be done to bring back that type of mass amazement and excitement towards technology?
When I think of the future I'm blown away by the possibilities of CRISPER, 3d printing, energy storage, block chains, driver-less cars, UBI, space exploration and more. We are advancing so quickly on so many fronts, and yet most people are indifferent or down right hostile toward it.
Things started to go wrong and limitations started to appear. The 1970s saw the US hit three big limitations: the failure of Vietnam, the oil crisis, and the collapse of Iran. On the technology front, that's when the environmental and social downsides of technology started to become really prominent. Three Mile Island, CFCs, leaded petrol, and so on.
That's why all the 50s and 60s stuff is so breathlessly unconditionally optimistic, and subsequent material has switched to skepticism.
People have become used to "new technology announcement === a few years later we're going to hear about the terrible downside, while the originators lied about all the safety risks". Driverless cars are absolutely in the middle of that at the moment. Vaping is currently experiencing a safety backlash.
The mid 70's were just about optimal. Lots of stuff you could pick apart and make new things from with standardized parts. After that more and more integrated circuits made it into consumer electronics and if the volume was large enough they'd be custom so useless to either repair or be used as a source of parts.
That’s an illusion. Could you, in the mid-’70s, repair a vacuum tube? A capacitor? No. You threw it away and got a new one, and you didn’t care how it worked inside. Older artisans were probably even then horrified by the ‘new’ age of people who didn’t know how to properly blow glass bulbs.
Maybe you couldn't repair the switch, potentiometer or the valve, but you could nearly always buy a standard replacement. From just about any suitable shop. Chances are the same switches and valves were used in dozens of products and makes, and there would be third party alternatives - cheaper with shorter life. Sure, there were a few uncommon specialist tubes in some things, but they were hugely outnumbered by standard components.
Now everything is specific and moulded to fit just one product. You try to buy a Sony XYZ1244/a switch, and find Sony don't make that part available, and it's custom moulded to only fit that one particular product. Maybe you can buy a switch as part of the whole control module that includes it, and a bunch of custom ICs at £99. Or maybe not as you're not an approved Sony repair shop.
When a part is botched it is botched, though - story time - I did more than once repair power transistors and valves with broken off legs. But that doesn't really matter, the idea was that parts were exchangeable between different manufacturers and that nowadays almost everything except a handful of passive components is custom made for that particular device. I understand the economics, but it does not help for serviceability or learning from what you are looking at. Good luck tracing the schematic of even a small IC, I spent hours taking apart color tv's and sketching the schematics of their guts to understand how it all worked.
that talks about a similar computer at MIT. I find all the pre transistor computation really interesting. Shannon was doing some really early work on what was, at the time, thought of as AI. Putting a mouse through a maze with movable walls. All with just circuits, no CPU. The circuits were not fixed, they had some branching based on phone switch tech (IIRC).
Sure, just like a sloppy slide-rule or mechanical calipers or micrometer will give you iffy readings. Whether the accuracy change is a material one depends on what you are doing. Daisy chaining results would get you into meaningless territory quite fast I would imagine.
Oddly enough, I have a circular slide rule where the printed scale is not quite concentric with the sliding discs, so the thing has a built-in periodic error.
Username checks out ;) That's neat, I've never had one of those but used a linear one for years for quick order-of-magnitude answers. I got to use it just as modern pocket calculators took off, the first ones were only four function jobs and they were priced insanely high so slide rules were more economic for the kind of problems I ran into in highschool. My maths book even had a whole chapter on using them. Then; after a few years I got a Ti-57 and that was that for the sliderule, that thing was programmable.
Not unlike the way floating point rounding errors accumulate, in fact. The only difference is you can mitigate those to a degree by increasing the digits of precision, whereas that can't really be done on a mechanical computer.
U.S. NAVY BASIC MECHANISMS OF FIRE CONTROL COMPUTERS MECHANICAL COMPUTER INSTRUCTIONAL FILM 27794 https://www.youtube.com/watch?v=gwf5mAlI7Ug