[kashprime]

ICE’s have over a century of development behind them now. I’m not going to miss them, but I wonder if there’s a way to ‘archive’ this kind of specialist mechanical engineering knowledge. It may come in handy one day.

It’s depressing to read about engineers having to go to the Smithsonian to study the lunar lander to relearn some of the innovations and improvisations made at the time.

[akamaka]

It seems inevitable that much of the practical knowledge is lost. Look at how few old sailing ships survive, despite that being the most advanced transportation technology for centuries. The schematics of the engines will survive, but future generations who study them will struggle to understand the design choices that went into them.

I’m happy that might change in the world of software, since so much collaboration is done online, and future historians will have troves of JIRA tickets to dig through.

[tablespoon]

> The schematics of the engines will survive, but future generations who study them will struggle to understand the design choices that went into them.

> I’m happy that might change in the world of software, since so much collaboration is done online, and future historians will have troves of JIRA tickets to dig through.

I doubt it. In my experience, JIRA tickets don't capture much information like that. Also, my employer at least has trouble maintaining actual design documents over the long term [1] that stuff that everyone views as transient has pretty much no chance.

Also, a lot of collaboration tools, while technically being more archiveable (e.g. slack, typical meeting recordings), are in practice so disorganized that they are rarely even useful as a long term reference for the teams that use them. Unless the team's software is truly world changing, I don't think any historians would bother.

Personally, I think the kinds of very formal processes that are unfashionable in the world of software are the ones that are most capable of capturing the reasoning behind design choices for future generations.

[1] stuff is spread across so many places, and a lot gets when we migrate from one enterprise collaboration system to another, or as teams reorganize.

[6pac3rings]

A team of six or eight should have a dedicated scribe to link everything that is going on in one coherent whole. That role functions as the interface to the team and onboards new team member when people move on. A team without a scribe creates a new opening for that role for a noob to learn by doing.

[jhanschoo]

> while technically being more archiveable (e.g. slack, typical meeting recordings), are in practice so disorganized

Think of all the papers this will produce for future historians!

[ehnto]

One advantage ICE has over the historical preservation of sail is that cars as a collectors item are far easier to keep around, and much of the design artifacts will last much longer.

It's unlikely ICE cars will be banned from being driven in the next 30 years, if ever. Once electric is proliferated the effects of car enthusiasts will be negligible. Even if there are some bans, motorsport will hopefully continue.

Many car manufacturers run heritage parts programs and there is already a well established industry for maintaining the collector and enthusiast car fleet that has nothing to do with the new car cycle. People are engineering entirely new parts for old cars all the time, we won't be forgetting how ICE works for a while yet. Although the goal of all that engineering effort may shift from efficiency to efficacy in the form of motorsports parts.

I fully expect to have to keep barrels of E85 at my house so I can take my soon to be heritage vehicle to the track every now and then, but I don't see us forgetting how to engineer for combustion engines.

[benrbray]

Will they? Most companies' internal systems are locked down and we'll suffer the same problem. Open source is the best way we have right now to preserve software knowledge.

[akamaka]

I’d bet the hosted services will hold onto data forever, so that they can mine it.

[devoutsalsa]

I don’t remember the lunar lander, but I do recall reading how each rocket engine on the Saturn V was one of a kind, because the engineers had to assemble each one by hand, and each acted slightly differently. So there’s no way to reproduce a specific Saturn V engine now, but there also never was back when the Saturn V was still flying.

[ColFrancis]

Genuine question: How is this different from a F1 engine where each piston head is crafted for each cylinder in the block? When you go high performance things that look like they're a commodity become one-of-a-kinds as you push them right to the limits.

[Aperocky]

Would we do any differently today? I think we gravitated towards smaller and more numerous engines (that are easier to construct to the same exact copies), but nobody since has built anything remotely as large as the Rocketdyne F1.

[sitharus]

These days we have computational fluid dynamics. We can simulate many more things to make an easy to build design rather than making manual adjustments to each one.

There’s been little need for a liquid fuelled engine that large though due to advances in solid rockets in the US and Europe, and a focus away from super-heavy launch vehicles elsewhere.

[xyzzy21]

You'd have to redesign it from scratch - it would be faster. Which is what happened with the SLS engine/launch vehicle.

Duplicating the F1 would involve technology and knowledge of individual engineers who are dead and corporations that no longer exist. The TACIT knowledge and technologies required to go with the blueprints died with them.

https://www.space.com/nasa-saturn-v-and-sls-compared.html

https://ourplnt.com/remake-rocketdyne-f-1-engine-humans-moon...

https://www.youtube.com/watch?v=ovD0aLdRUs0

[cameldrv]

The RD-170 is about the same as an F1 although it has 4 nozzles and 4 combustion chambers.

[samus]

SpaceX pursues its business on a much larger and, most importantly, more regular schedule. In this case it is possible to standardize and mass produce many parts.

[xyzzy21]

No there isn't.

Technology knowledge is extremely ephemeral and easily lost once you stop using it because so much of it is "tacit knowledge" that is only preserved by active use and passed by mentorship/apprenticeship.

We've lost many technologies this way. In some cases like FOGBANK and the recent B2 bomber heat exchange system, there was enough incentive to reconstruct it from scratch at great cost. For most lost technologies, that incentive will never exist. If you are very, very lucky, some other country will maintain it by continuing to manufacture it.

This is a major reason why "bringing manufacturing back to the US" is so hard. We literally lack the technology/knowledge context as a society to do that most of the time. It's like pondering a semiconductor fab project in Chad or Sudan!

Otherwise it becomes like pyramids and monoliths that some "moderns" can't imagine came from anything other than space aliens! "They couldn't have made anything like that back then because we can't do it today."

[dv_dt]

Well yes and no. Sometimes you can bring the industry back, but with a newer generation of technology. High capital investment industries like manufacturing I think are particularly amenable to generational leapfrogging because the original capital investors are very tempted to keep operating old methods for the margin. This gives an opening for any entity willing to make a new capital investment.

This is essentially what China did to a lot of US manufacturing, as well as SpaceX to old line aerospace giants.

[tra3]

Don’t even have to go back that far. US govt is calling for contractors to reverse engineer b2 bombers: https://news.ycombinator.com/item?id=26333649

[mhh__]

A B2 bomber and a completely unclassified car engine is a slightly different order of magnitude in terms of complexity.

If you want to run a current F1 car in 40 years time then you will need to RE it however because the Powertrain is complicated and totally custom

[jjav]

> If you want to run a current F1 car in 40 years

Never mind 40 years, F1 cars of even a couple years ago become impossible to run because just getting the engine started requires custom equipment and knowledge that often doesn't seem to be preserved year to year.

[dvdkhlng]

I think we have already "lost" similar amounts of knowledge when the mechanical typewriters went away. And the mechanical calculators (and mechanical clocks?).

Or have you recently looked through a book on vacuum tube circuit design? Even more modern transistor-based analog circuit design is going to fade somewhat, as more and more parts of devices are moved to digital circuits.

It is a little terrifying and sad to see how many incredible ideas, that took a lot of smart people to develop, are basically thrown away over time.

[pmastela]

For posterity, I wish all this information would make its way to a public repo as to not be lost in the sands of time.

[samus]

What purpose should such a repo serve? The most common ones are museums. Yes, they are not perfect, and it is easy to oversimplify and misinterpret what is presented there.

If you want to preserve it for survivors of civilisational collapse, then it has to be codified and simplified first. Well done, a civilization could jump tech levels much quicker than if they had had to bootstrap everything first.

[burlesona]

I know what you mean, but at the same time that’s just how tacit knowledge works. How many times have you looked back at software you wrote 4+ years ago. There are lines of code in there that you can’t remember the reason for that look strange or out of place, but you have a vague sense that they do something important. The actual knowledge is gone and the only way to rediscover it is via re-working the problem.

If we stop making ICE cars then a lot of the tech around them will be lost, but chances are we won’t care a whole lot because electric cars will be better for most people in most ways, and ICE cars will live on as a historical curiosity.

[jvanderbot]

I just paid a blacksmith to teach me to forge weld. Someday maybe bored middle-class folks will pay to learn to change oil on a v8 between marathons.

[rasz]

Brazing? like https://www.youtube.com/watch?v=pz9H850dtqY ?

[jvanderbot]

Nope: like this: https://www.youtube.com/watch?v=j814AqiKVvE

[rasz]

Ah, so its purely from a historical preservation angle. Is there a point with $80 Amazon stick welders available?

[Bancakes]

That's amazing. Do you have any more such channels, on rebuilding after an apocalypse?

What about books on skillsets necessary?

[skissane]

> ICE’s have over a century of development behind them now. I’m not going to miss them

They will mostly disappear from cars, and (more slowly) even trucks, but they are still going to be around for a lot longer in other applications – backup generators, heavy construction and mining equipment, the military, locomotives, smaller aircraft, ships and boats, etc.

We may well start to lose some knowledge about how to design ICEs optimised for specific use cases, we aren't anywhere near forgetting how to design ICEs in the general case.

[axaxs]

Yeah, it's always a real shame when that happens. We have the internet, now, at least, hopefully lots can and will be made available once ICE falls further out of favor.

I'd read recently something similar about Colt. Not only did they 'lose' their bluing formula, but apparently never documented much about their smithing process, as they had trouble hand fitting revolvers due to a knowledge gap.

[0xfaded]

At the front of my mind are aircraft piston engines, which haven't progressed much beyond the 60s. One notable exception are the Diamond Astro engines, which are based on mercedes blocks.

Seeing as mass scale ICE development is probably a deadend, hopefully the current state of engines will hold over until battery energy densities become viable for light aircraft (an extremely selfish desire, i admit).

[rasz]

Lack of civilian aviation development is a result of misguided regulation. "OMG stuff is dangerous, lets require expensive certification for even the most trivial things ... except for the dangerous stuff that is already in use right now" and then we end up with brand new 2021 Helicopters shipping with 70 year old Lycoming engines.

[sjwright]

Even if we get to 99% EV within fifty years, it’s going to be a long while after that before EV takes over 100.000% of ICE.

[pmorici]

There are something like 1.4 billion vehicles in operation world wide and 80 million new vehicles are sold every year.

That is at most 17 years once new vehicle sales are effectively 100% electric which will probably happen around 2030. The value of used gas cars will deteriorate rapidly as the spare parts and repair networks disappear and electric vehicle costs become much cheaper pulling forward demand for the transition to electric.

[donavanm]

You making a lot of assumptions that arent born out by current experience. Its not uncommon for 100 year old cars to be registered and driven on public roads today. I personally have registered “daily drivers” from the 1970s and 80s. I actually rebuilt my 1972 motorcycle using factory parts from the dealership. As cars get rare they tend to increase in value, which also increases the return on effort/money put in to them. The “last running ICE car” is going to be a prized gem in a collection, not scrap metal.

[jeppesen-io]

and when 90+ precent of the cars on the road are EV - where are you going to get gasoline? or think how much an oil change will cost when motor oil becomes a specialty item

> You making a lot of assumptions that arent born out by current experience

current experience assume easy access to gasoline and motor oil that has been true for 100 years - that will pass in time

[flukus]

I've been wondering if the cutoff is actually much earlier. Say 50% of cars on the road are EV, how many gas stations can continue to run profitably with 50% less customers? Even getting and refining the gas requires massive global supply chains to transport and refine, what happens when there's a 50% drop in customers? Do they lose their efficiencies of scale? Do prices rise substantially as this happens?

I suspect once we reach a certain thresholds it will create other push factors toward EV's and accelerate their adoption.

[loonster]

It is my understanding that refineries cannot change the percentage of gasoline that comes from a barrel of oil. If electric cars become as popular as ICE, either gasoline will be used in other things, or other oil distillates will become very expensive. Likely no more cheap air travel, propane, etc.

I expect we will hit a wall that will inhibit adoption.

[burlesona]

Gas stations are an interesting case, because most of them make most of their money selling snacks and beer. The gas money largely goes to the oil company.

[hef19898]

I have one, too. Not really as a daily driver, gas mileage just sucks. When the daily driver is about to be replaced one day, it will certainly be by an EV.

I think that the main issue will be maintenance of ICEs in the future. Like try to find a garage that is able to tube carburators these days. It will be a while until we reach that point so.

[rasz]

You can DIY mod carb engine to EFI using plethora of cheap user programmable ECU modules and oem/salvage yard parts, its almost Lego at this point in time with everything understood and tons of documentation in the open. Throttle body, MAF, Lambda sensor, one fuel injector, fuel pressure reg, fuel pump, crankshaft position sensor, ECU, laptop to program and tune the ECU, all in all <$500. As a bonus you get to delete distributor and convert to electronic ignition for free. There are even plug&play kits for early troublesome mechanical fuel injection systems in ~$600-1000 range https://kjetkillers.pl/en_US/c/EFI-Electronic-Fuel-Injection...

You dont need top of the line $1700 Haltech elite 2500 or Link G4X FuryX if you arent seriously racing.

~$350 full standalone ecu http://ottomotive.pl/sklep/komputery-standalone/easyecu-2-pl... or http://ecu-shop.eu/produkt/ottomotive-easyecu-micro/

~$150 simple piggyback able to work in standalone mode http://ottomotive.pl/sklep/komputery-piggyback/digital-ecu-c...

[hef19898]

Maybe one day! I also have the option to take the EFI from one of the following model years. I suck at electronics, and the wiring is already a royal mess (wasn't me so!).

[donavanm]

Exactly why I brought up very early cars and my carb’d motorcycle. They go out of common use but dont disappear anymore than carriages or horses have. Fully urbanised and developed countries have them as a collection or passion project and do silly things like bodge or manufacture specialty parts and fuel blends while reading esoteric haynes manuals on carb tuning. Meanwhile they remain commonplace where it makes sense, small portable two strokes, aircraft, developing nations, etc.

The thought that civilization is going complete displace and lose hydrocarbon based engines and logistics in 17 years... Its optimistic I guess?

[hef19898]

Well, you still get OEM parts for Zenith carbs from the 60s. So I guess there will always be a rest of ICE powered cars. I wouldn't mind using green fuel or gas. Which might be the way forward for classics, mainstream I honestly only see EVs and fuel cells in the developed world.

[pmorici]

Museum pieces are not what we are talking about here.

[onion2k]

That would be true if the oldest cars are being removed from the pool, but that isn't how it works. Cars crash, some models wear out, some just have too little resale value to keep going, etc so they're removed from the pool earlier. Other cars are very desirable so they're get around much longer. It's not a perfect circular system.

[sjwright]

I was talking about the proportion of the worldwide fleet, not proportion of new sales. Remember, not all vehicles are passenger cars. And not all vehicles always near electric infrastructure.

[p1necone]

I don't know about that, consider other tech that got superseded by new methods. It only took a couple of decades for there to be no companies manufacturing CRT TVs any more.

[sjwright]

The use cases of vehicles are much wider than the use cases of TVs.

[jeppesen-io]

fifty years? no way. But even if you're right, what point are you trying to make?

[sjwright]

My point is that there's a long tail of vehicles out there. Whether the fleet transition from ICE to EV is fast or slow, getting the long tail fully transitioned in a similar timeframe is implausible.

Heck, we haven't even transitioned 100.000% away from horses yet.