General Physics Thread

>>8707883
>It's a well known fact that light travels in the path that takes the shortest amount of time, not necessarily the shortest distance

https://www.youtube.com/watch?v=-WHpQVMZbjo

>>8707908
https://en.wikipedia.org/wiki/Fermat's_principle

>>8707883
No. Light travels in straight lines. Also even if it were possible for light to move like you describe then it would need to imply information propagating backwards in time (how would the photon know to bend before it hit the object?) I think there are some strange metamaterials that cause some strange effects, but nothing like you described. Finally light travels the path such that it's optical path length and geometric path length are both minimised (Fermats principle), the bending nature of the path would seem to imply that the geometric path length isn't minimised.

There, three reasons this wouldn't happen irl.

My turn:

I'm trying to do this classical mechanics question:
>Assume that the earth consists of a core of uniform density [math] \rho _c [/math] and a mantle of density [math] \rho _m [/math]. Further assume that the boundary between the two is of a similar shape to the outer surface, only with a radius 3/5 as large. Find the ratio [math] \rho _c / \rho _m [/math] that explains the observed quadrupole moment.

I've been at it the past couple of hours and I'm at a bit of a loss.

File: g1.png (164KB, 600x600px) Image search: [Google]

164KB, 600x600px

you can try it yourself with a laserpointer, but the effect is minimal
https://en.wikipedia.org/wiki/Knife-edge_effect
If instead there was a smooth gradient of optical density, then it would act just like a regular lens

File: 220px-Snells_law2.svg.png (15KB, 220x395px) Image search: [Google]

15KB, 220x395px

>>8707916
>Also even if it were possible for light to move like you describe then it would need to imply information propagating backwards in time (how would the photon know to bend before it hit the object?)
But this very thing happens with 2 different media. How does the photon know what angle to travel at when going from P to Q? It just does, that's the mystery of Fermat's principle.

>the bending nature of the path would seem to imply that the geometric path length isn't minimised.
It is if the index of refraction of the material is so large that the bent path actually takes less time then a straight path through the material.

>>explains the observed quadrupole moment.
If you explain what this is then I'll try to help.

>>8707921
>but the effect is minimal
But what if the object we are talking about is a galaxy? Then the effect would be amplified I'm guessing.

>>8707916
This all makes far more sense when you realize there are no photons, that light does not "travel" at all, and that the "speed of light" is nothing more than a rate of induction through the medium.

>herp derp we so smart now

>>8707945
>How does the photon know what angle to travel at when going from P to Q?
You don't tell a photon in P "go to Q", you just send it off in some direction and it ends up at Q.

>>8707883
Never going to happen.

If it did, complex optical bench setups would just fail because instead of following the designated optical path the light would just skip the entire thing and go straight to the detector. In the last hundred years of optical experiments this has yet to happen.

>>8707984
Also just to add, its better to think of light travelling in the path that takes least time as being a result of Snell's law, not the other way round. Snell's law is really just conservation of momentum.

>>8707945
>But this very thing happens with 2 different media

No. There it hits the boundary and slows down. In your situation it would have to curve out of the way before it hit the boundary. Implying the knowledge of the refractive index reached the photon before the photon and boundary interacted.

>It is if the index of refraction of the material is so large that the bent path actually takes less time then a straight path through the material.

Again no. I got my statement Fermats principle wrong (although the result remains unchanged) the opitical path length is the geometric path length (which is just a distance in R^2) times the refractive index. Which would imply for any constant index the geometric path length would still be straight lines.

>>8707955
>>8707984
>>8707994
I guess this makes sense. I must have misunderstood Fermat's principle.

>>8708010
>There it hits the boundary and slows down. In your situation it would have to curve out of the way before it hit the boundary. Implying the knowledge of the refractive index reached the photon before the photon and boundary interacted.
This all works until you consider the photon's POV, where there is no "before it hits the boundary". In any case, I guess I wasn't considering that light travels in a straight line in uniform medium. But I seem to recall that light travels every possible path, and the "actual" path it takes is just the average of all these.
https://www.youtube.com/watch?v=kMSgE62S6oo

>>8707916
Indeed, light travels at straight lines in space. But what of the whole space curves.

>>8708052
>This all works until you consider the photon's POV, where there is no "before it hits the boundary"

There is no frame where light is at rest, so we can't talk about "what the photon sees".

>>8708065
Then it travels along a geodesic.

>>8707954
oh it's you again
i'm sending the assassins