"Old tech" often just works. Flawlessly. I have a dumper truck I borrow from next door. Crank handle, diesel. It just works. Starts every time. No battery, no crap to go wrong.
"new tech" is often designed to last 5 or 10 years maximum. After that is anyones guess as to whether it'll work.
http://www.telegraph.co.uk/motoring/green-motoring/11187483/...
Think of all those cars that didn't get built because I drove the same car for 25 years. I thought about that often.
The only time it didn't pass emissions test is the gas cap didn't seal properly. Bought a five dollar cap and was good to go.
And it gets 35 mph in daily commuting.
I loved that thing. If you turned on the original heating system, you'd die of carbon monoxide poisoning, but I loved it.
Some old tech doesn't work flawlessly. Steam power vehicles, like locomotives, cars, and tractors, require more maintenance than their new tech replacements.
Yeah there's some survivor bias, but the materials used to make things in the olden days was just better material. On the down side, sewing machines used to cost a months wages or something.
But not all materials are better. Celluloid is the first thermoplastic, and the first material used for movie films, but it's highly flammable and has been replaced by acetate film. Celluloid was also used to replace ivory in billiard balls, which was better than the clay and wood balls used earlier. All of those materials are worse for professional play than modern billiard balls made from composite plastics.
You brought up sewing machine. Sewing machines were made in the hundreds of millions. What you see are the ones that weren't trash and were maintained. Here's a page complaining about people complaining about people selling old sewing machines that mostly useful as "boat anchors" http://runningstitches-mkb.blogspot.se/2012/08/open-letter-t... . Quoting one example, "I'm not impressed when you tell me your 1950 Singer has "only been used twice to sew on patches" (note: I am not making this up!) If this machine has been sitting unused in an attic for half a century, it is most likely frozen or seized up."
Quite clearly a lot of old sewing machines do not "still work 50 or 60 years later."
A Singer 201, which is widely considered the classic sewing machine, cost 6 months wages, not the 1 month you think it was. See http://www.sewalot.com/singer_201k_sewalot.htm . (Also, Singer switched from cast iron to aluminum a bit over 60 years ago, so your timeline is a bit off.) There are a lot of crappy modern sewing machines in the <1 week's salary range. But that's hardly a fair comparison, is it? Plus, how many of those have an original motor that's been in use for 60+ years without wearing out?
My Mom has a cast iron Singer 201. Which she loves even though it only does straight stitches. She also has several other sewing machines. She still uses her Bernina from the 1980s. So at least some modern (less that 50 years old) machines can and do last longer than 5 years.
Finally, there are plenty of sewing machines on eBay from the 1950s and early, for cheap. What is the reason that they aren't scooped up if they are of significantly higher quality than modern machines?
Incidentally, mechanical diesels are also immune to EMP. All they need is fuel and air. Not true of newer electronically controlled ones, which is done mainly to reduce emissions.
The compression-ignition characteristic also means diesels have this rather spectacular failure mode: https://en.wikipedia.org/wiki/Diesel_engine_runaway
Extreme example: I once read a fascinating article about middle aged technology, where they commented on, among many things, how a high-quality cart wheel would have a hard, dense wood on the axis, another kind of wood in the spokes that was flexible, and a border made of a very stiff wood. I'm searching for it but I don't seem to be able to find it, alas.
And anyway, why are we still teaching people how a CPU without branch prediction works? When will we let go of old technologies?
[1] https://en.wikipedia.org/wiki/Carburetor#History_and_develop...
Fuel injection systems have replaced carburators. Branch predictors haven't replaced anything, if you want to compare them with any engine that would be turbo chargers (i.e. they're optional).
I'm not sure comparing CPUs with engines makes a lot of sense though.
You should really improve your metaphors.
My dad gave me some good advice years ago. He told me to buy a year specific service manual for every automobile you ever buy.
Here in Europe, it's not uncommon to see cars in good shape which have over 400k miles on the clock. If you think about it, a European or Japanese mid 90's car with no rust and a rebuilt engine will give you more trouble-free miles, for a fraction of the money.
The average car on US roads is 11.4 years old: http://www.latimes.com/business/autos/la-fi-hy-ihs-automotiv...
The average car on EU roads is 8.6 years old. In fact, in the EU only Estonia and Finland keep their cars longer than the US on average. http://www.acea.be/statistics/tag/category/average-vehicle-a...
I have worked on thousands of carburetors. I can re-jet the 2 carbs in my race bike in about 5 minutes (main jets only). Carburetors are indeed in use on millions of engines. I understand Bernoulli's principle, and I also understand that a carburetor is lousy at part throttle metering. What does that mean to you?
Part throttle is anything that is not full throttle. Some people have a hard time believing that a fuel injected engine offers no full throttle advantage in total horsepower - it is the barely open part that matters. Fuel injection is massively more efficient and clean than carbs at part throttle. The part throttle issues cause all the emissions problems, cause all the pollution, cause all the smell when following behind a carb'd car.
The electronic carbs of the 80's were a real nightmare. Take a mechanical device, put some dozen vacuum lines on it, each controlled by a solenoid, give it a dumb computer that you can't talk to, and no tech information anywhere. Not even at the dealer. Caveman tech, dressed up with bunch of patches and cruft. It was done to save a buck, not because the better way wasn't known. And the systems are horrible and should be sent to the crusher. It is fuel metering in the worst way; adding miles and wear just makes all the problems even worse,
The EEC-IV systems used by Ford from 87-95 only had a 60 pin connector on them, the same number of pins that a Beagle Bone Black has. I have be thinking about how to use that as a basis to replace existing systems with an open source system that would be better and cheaper.
But carburetors need to die. They are not great new, and get worse with age.