3 ms·

Knowing Newtonian physics doesn't mean we understand all things that move. The point about "not understanding" flight is that, if we truly understood it, we co

Knowing Newtonian physics doesn't mean we understand all things that move. The point about "not understanding" flight is that, if we truly understood it, we could design the optimal aircraft from first principles before it ever entered a wind tunnel. Instead, we work based on incrementally improving tribal knowledge of what has worked in the past and try to make something similar to fit our desired flight envelope. Compared to something like rocket science where the entire craft can be built on a computer and we'll know exactly how much cargo we can get to the moon without even turning a single screw, we don't understand flight.

garmaine 5y ago
I don’t think this is true. There are many physical ststems for which we know the underlying physics very well, but the equations can’t be simply solved, and numerical simulation is more costly than just building the damn thing and testing it. Wing lift under turbulent conditions is one of those things. So we use wind tunnels. Not because we don’t understand lift—we do—but because it’s just easier. This is getting less and less true with each generation of supercomputers though. EDIT: There is perhaps a better way of explaining it for this crowd though. To use numerical modeling to predict performance is to take a physical problem and turn it into a computational problem. And while engineers understand physical systems pretty damn well, us computer scientists have largely failed at the objective of making software systems with hard reliability guarantees. You can write a fluid dynamics simulation to test your new wing design, but how do you know that the simulation does what you think it does? Even if the code has been tested before, how do you know you're not now hitting some sort of edge case? At the end of the day, you have to build the damn thing to test it. Numerical simulation are used more and more these days as the codes are refined, computers get more powerful, and engineers have more trust in their capabilities. But traditionally, and still a lot of the time, they build prototypes and test in wind tunnels because reality never fails to model physics accurately.
- meheleventyone 5y ago
  The edges really show for smaller craft. At low Reynolds numbers lift gets funky. A lot of flight research is around making very small things fly efficiently.
jessedhillon 5y ago
Almost nobody understands software then, by this definition.
- Nevermark 5y ago
  Software is definitely still a craft, not engineering in the build-a-bridge sense. So I would say, nobody really understands software, but many people have experience and total experience is growing. Via new languages, algorithms, patterns, etc. The rapid experience advancement suggests there is a lot unknown and not understood.
- foobiekr 5y ago
  I can’t tell if this is meant as a counter or not. Literally the pool of people who really understand capital S Software is tiny. I’ve worked with thousands who don’t and maybe two dozen who do.
Cogito 5y ago
I'm not sure if this view undersells aero or rocket engineers more. Rocket engineering uses an immense amount of both modelling and physical testing. No-one says "Well I've got the Tsiolkovsky rocket equation, so let's go to the moon!" I'm not even sure what the bar of 'understanding' is here - the fact that we have iterated and improved on powered flight as much as we have necessarily means that we understand it on a deep level, let alone the fact that we can create excellent models that predict what will happen to a wing in different situations. At what point would you say we do understand flight?
krisoft 5y ago
You have an unduly rosy picture of rocket science. The turbopump used in most liqued fueled rockets to presurize the fuel is notoriously complicated. Small changes to the design can result in a dramatic loss of efficiency, or even worse if it starts cavitating the pump can eat itself. For this reason nearly nobody designs turbopumps from scratch and first principles, they take a well understood design and maybe tweek it a bit. And you can bet that they then test those tweeks on a bench a lot. Similarly devilishly complicated is the injector design. Obviously you want to mix the oxidizer and fuel in the optimal ratio for the highest efficiency. That’s the easy part. But then you also want to offset from this optimum near the edges to produce a colder flow near the nozzle wall to protect it from melting. Of course nowadays people do a lot of computer simulations to save on testing time, but it is still not uncommon to discover combustion instabilities or hot-spots in the engine tests. So no, nobody can, let alone did, design a rocket entirely in a computer and then send it to the moon without many many tests, and incrementally improved tribal knowledge.
- foobiekr 5y ago
  I can’t think of a technology sector where this logic isn’t true. I also can’t think of a technology sector where the naive/new/low level/clueless don’t also assume this is not true. Even manufacturing. ASICs. Software. Everything.

garmaine 5y ago

I don’t think this is true. There are many physical ststems for which we know the underlying physics very well, but the equations can’t be simply solved, and numerical simulation is more costly than just building the damn thing and testing it. Wing lift under turbulent conditions is one of those things. So we use wind tunnels. Not because we don’t understand lift—we do—but because it’s just easier. This is getting less and less true with each generation of supercomputers though. EDIT: There is perhaps a better way of explaining it for this crowd though. To use numerical modeling to predict performance is to take a physical problem and turn it into a computational problem. And while engineers understand physical systems pretty damn well, us computer scientists have largely failed at the objective of making software systems with hard reliability guarantees. You can write a fluid dynamics simulation to test your new wing design, but how do you know that the simulation does what you think it does? Even if the code has been tested before, how do you know you're not now hitting some sort of edge case? At the end of the day, you have to build the damn thing to test it. Numerical simulation are used more and more these days as the codes are refined, computers get more powerful, and engineers have more trust in their capabilities. But traditionally, and still a lot of the time, they build prototypes and test in wind tunnels because reality never fails to model physics accurately.

meheleventyone 5y ago
The edges really show for smaller craft. At low Reynolds numbers lift gets funky. A lot of flight research is around making very small things fly efficiently.

jessedhillon 5y ago

Almost nobody understands software then, by this definition.

Nevermark 5y ago
Software is definitely still a craft, not engineering in the build-a-bridge sense. So I would say, nobody really understands software, but many people have experience and total experience is growing. Via new languages, algorithms, patterns, etc. The rapid experience advancement suggests there is a lot unknown and not understood.
foobiekr 5y ago
I can’t tell if this is meant as a counter or not. Literally the pool of people who really understand capital S Software is tiny. I’ve worked with thousands who don’t and maybe two dozen who do.

Cogito 5y ago

I'm not sure if this view undersells aero or rocket engineers more. Rocket engineering uses an immense amount of both modelling and physical testing. No-one says "Well I've got the Tsiolkovsky rocket equation, so let's go to the moon!" I'm not even sure what the bar of 'understanding' is here - the fact that we have iterated and improved on powered flight as much as we have necessarily means that we understand it on a deep level, let alone the fact that we can create excellent models that predict what will happen to a wing in different situations. At what point would you say we do understand flight?

krisoft 5y ago

You have an unduly rosy picture of rocket science. The turbopump used in most liqued fueled rockets to presurize the fuel is notoriously complicated. Small changes to the design can result in a dramatic loss of efficiency, or even worse if it starts cavitating the pump can eat itself. For this reason nearly nobody designs turbopumps from scratch and first principles, they take a well understood design and maybe tweek it a bit. And you can bet that they then test those tweeks on a bench a lot. Similarly devilishly complicated is the injector design. Obviously you want to mix the oxidizer and fuel in the optimal ratio for the highest efficiency. That’s the easy part. But then you also want to offset from this optimum near the edges to produce a colder flow near the nozzle wall to protect it from melting. Of course nowadays people do a lot of computer simulations to save on testing time, but it is still not uncommon to discover combustion instabilities or hot-spots in the engine tests. So no, nobody can, let alone did, design a rocket entirely in a computer and then send it to the moon without many many tests, and incrementally improved tribal knowledge.

foobiekr 5y ago
I can’t think of a technology sector where this logic isn’t true. I also can’t think of a technology sector where the naive/new/low level/clueless don’t also assume this is not true. Even manufacturing. ASICs. Software. Everything.