[soVeryTired]

I've heard people make the argument about GR being essential to the GPS system before, but the skeptic in me has doubts. I don't deny that it's useful to have a correction derived from relativity, but is it really as essential as people make it out to be? If we hadn't known about GR when we built the GPS system, wouldn't we have realised there was an unexpected drift in the clocks and found a way to recalibrate them periodically?

[mitchty]

Yes its essential. Because the satellites are moving at orbital velocity, their frame of time is ever so slightly different than ours. To keep accuracy timewise without special and general relativity would be difficult if not impossible without understanding relativity generally. Could we have done GPS without relativity? Maybe, but we'd have to discover it with the satellites to be able to use it.

A quick note from: (i'm too lazy to do the calculations right now) http://www.astronomy.ohio-state.edu/~pogge/Ast162/Unit5/gps....

    If these effects were not properly taken into account, a navigational fix based on the GPS constellation would be false after only 2 minutes, and errors in global positions would continue to accumulate at a rate of about 10 kilometers each day! The whole system would be utterly worthless for navigation in a very short time.

[hinkley]

I think I see what he's saying. Those satellites are moving at a fixed velocity. So the time dialation effect on them relative to the ground should be static.

It could reasonably be the case that we just adjusted the clocks on the satellites by .05% or whatever the drift rate is by measuring them against known points until we get it right. Good enough engineering.

Eventually someone would have asked some scientists to explain why it's happening.

[mitchty]

> I think I see what he's saying. Those satellites are moving at a fixed velocity. So the time dialation effect on them relative to the ground should be static.

The issue you have is that the dilation effects aren't static, they're all relative to each satellite and ground observer and are constantly shifting based on the orbits. Basically the premise you have to accept to allow for "just adjust the clocks" is too basic. This is why both general and special relativity come into play in GPS. You might get away with adjusting for a single observer, but not all observers.

Note that we already have the clocks purposefully skewed to account timewise for their orbital speed and still require constant updates. I just don't see that happening through "good enough engineering". If you're a nanosecond off you're off by over 1 kilometer and things get worse from there.

Plausible as a hypothetical gedanken experiment? I suppose, plausible in reality? I'm skeptical that you'd be able to do it. Would be akin to launching a rocket to the moon without understanding how to fly.

[soVeryTired]

Nice, thanks for the info! I'd never seen the inaccuracy quantified like that before.

[beojan]

> If we hadn't known about GR when we built the GPS system, wouldn't we have realised there was an unexpected drift in the clocks and found a way to recalibrate them periodically?

And hence have discovered GR in the process. Alternatively (and more likely) we would have decided the idea simply doesn't work.

[jerf]

"Alternatively (and more likely) we would have decided the idea simply doesn't work."

No, we would just keep beating on it until it works. It's a myth that science always precedes engineering. It's a very regular drift, and worst case scenario, if they just couldn't work it out, they'd build several ground stations in known locations, derive corrections as those drift, and upload corrections into the system periodically. (Similar things are done today: https://en.wikipedia.org/wiki/Differential_GPS )

Even with the science that we have, there's still fudge factors and empirically-determined values we have to use anyhow, because the Earth is not homogeneous and we have to take its slightly-lumpy gravity field as a given, not something we can "scientifically" determine. (Of course we "use science" to determine the lumps, but the lumps themselves are simply a given.)

[Newtopian]

Given the choice between trial and error engineering and science backed engineering I much prefer the latter. That does not mean doing engineering without the science is impossible just so much harder. They both go hand in hand, nothing mythological here. Advancements in one leads to ideas in the other and vice-versa.

[jerf]

"Given the choice between trial and error engineering and science backed engineering I much prefer the latter."

That's not the question; the question is, are we forced to give up engineering if we don't have "the science" yet?

And the answer is an objective "no", from abundant past human history. There's this myth sold that science always precedes engineering that is very, very popular. I'm not even sure where it's coming from. Oral history in primary education, maybe. But it's a myth that engineers themselves can ill afford.

The vast majority of practical programming is programming running way, way ahead of the "science", which occasionally takes point samples of how 10 college sophomores behave under a certain limited experiment.

[hinkley]

We just overbuild until we have the science and the manufacturing capacity to back it up. Every once in a while I hear some civil engineering fan comment about how the roman aqueducts are still standing because they built a huge margin of safety into them (because they couldn't do the calculations, not because they wanted them to last a thousand years)

[mnl]

Maybe it's because for instance there were exactly zero engineers working with radio waves before Maxwell and Hertz, or zero engineers working with semiconductor transistors before, basically, Brattain, Bardeen and Shockley. Electromagnetism and electronics happen to come straight from science labs. You could check also Idk... chemical synthesis, polymer science... nuclear physics?

I don't understand why would you come up with such "objective" statements unless you really think it's not necessary to know anything about the history of science and engineering to have a strong opinion about them.

[littlestymaar]

> Maybe it's because for instance there were exactly zero engineers working with radio waves before Maxwell and Hertz

Well, humans have been crafting optical lens way before Maxwell ;). And even way before [1] Descartes and Newton decided to study light.

For everything out of reach of our senses, like the examples you gave, we need formal science. But for everything humans can see, smell or touch, we're pretty good with empirical observations : chemistry, fluide dynamics, genetics and mechanics where comonly used way before formal science was even a thing.

[1]: https://en.m.wikipedia.org/wiki/Nimrud_lens

[jononor]

Isnt the more relevant question "does engineering benefit from (basic) science?"

Possibly amended to include "to a degree that justifies the spending"

[JackFr]

> And hence have discovered GR in the process.

Fascinating to wonder how modern physics community worldview would be different, if GR had gone [empirical anomaly -> new theory], and ended up in the same place rather than [thought experiment -> empirical proof]. (IAN a historian of physics - forgive me if I have that wrong.)

> Alternatively (and more likely) we would have decided the idea simply doesn't work.

I don't think that's true. Consider the Bell Labs guys who won the Nobel Price for Microwave Background Radiation, who were just trying to remove the noise their antenna was receiving.

[smueller1234]

Bell Labs and the CMB is nice example! At the same time, it unfortunately doesn't give us any intuition about how many times somebody doing engineering had less tenacity or experience and gave up instead of ending up with a Nobel Prize. :)

[dragonwriter]

If you don't have GR, you aren't going to do gravity maps (because you have no reason to think you need them).

Without that, you don't have much opportunity to discover the basis of the correction (which, as others have pointed out, is just discovering GR, anyway.)

[dekhn]

I think it's reasonable to believe that if we hadn't found GR, and implemented GPS, that observant scientists would have found GR.

The reason is that for econmics reasons, GPS developers work hard on noise reduction. Once you eliminate all the "technical" forms of noise, you're left with the "scientific noise".

[carapace]

> an unexpected drift in the clocks

In every kind of clock.

I would hope in this scenario that that would have been noticed and it would have aroused sufficient curiosity to investigate and eventually discover that the effect was due to time itself flowing at a different rate.

[petters]

Imagine trying to find that bug.

Finally, after an enormous effort being able to tell your manager that the error is in the space-time of the universe.

[carapace]

Riiight?

It reminds me of something I read about the search for what we would now call "violations of conservation of energy". Scientists kept coming up with more experiments to try to show this phenomenon, but none of them worked. Eventually they were forced to conclude that that's just the way it is:

"A perfect conspiracy is a law of nature."