4 ms·
Theoretical physics is infested with cartoons. Squiggles. Little arrows. Hand-drawn scribbles. I've looked into why there's a total vacuum of numerical simula
by jiggawatts 6y ago
Theoretical physics is infested with cartoons.
Squiggles. Little arrows. Hand-drawn scribbles.
I've looked into why there's a total vacuum of numerical simulation, and the reason is simply snobbery. There's a philosophy that numerical methods are what mere engineers do, and not as intellectually pure as symbolic solutions.
You regularly hear comments like: "Such and such cannot be solved", what is often meant is: "It can only be solved numerically, but not with closed form algebraic expression."
People that say things like the former statement have excluded numerical solutions as valid in their mind. Such solutions may as well not exist, as far as they're concerned.
I see this attitude everywhere in theoretical physics.
The other issue is that because of the seventeen layers of assumptions and abstractions, most physicists are now at the point where they're not really making theories about physics, but instead they're debating the properties of the abstractions themselves. It's like kids trading basketball cards and saying one player is better than the other because his card is printed better.
This can all be boiled down to an acid test: Can you render your equations? As in, full 3D numerical simulation with an image as an output? Not a graph, not a scalar value, but a picture?
For simple (non-diffracting) optics? Yes! That's literally every 3D computer renderer!
For classical electromagnetics the answer is: yes.
For plasma physics the answer is: yes.
Even for special and general relativity, the answer is: yes.
For QED? Err... maybe? I've seen some toy examples in 2D, and I think I've seen a 3D example once. In principle, it's doable.
For anything else? Nothing. Or at any rate, I've seen nothing despite years of searching.
E.g.: Can you show me any "rendering" that extends a QED simulation with the weak interaction?
Can you find any animated examples of any particle interaction? E.g.: a free electron being captured by a proton, or an electron changing orbitals?
Again: Images or animations please. Not graphs or scalars. Everyone I ever challenge gets confused and links me to a paper with a histogram in it. That's not what I mean!
- drdeca 6y agoNot sure what criteria / desiderata you have for a picture. Like, how do you want the different possible outcomes to be depicted together? Are you asking for a picture of 2 spheres and then another sphere? What are asking for?
- jiggawatts 6y agoIsosurfaces? False colours? Volumetric rendering? Cross-sections showing field strength through that plane? Field strength at solid surfaces? Pick any one. I find even questions like this revealing: Why are you asking me such incredibly basic questions instead of having a dozen renderings ready to hyperlink? If you ask a NASA guy to link some airflow simulations, he won't retort with "what criteria would satisfy you?". He'll just straight up link a bunch of gorgeous renderings of physically accurate airflow simulations. Not cartoons.
- drdeca 6y agoWell, I’m not a particle physics guy (or a physics person in general. I’m more a math guy.), so it’s not as bad as if someone really in the field responded as I did. I guess my confusion (which may be from ignorance of physics) is about how to handle depicting superpositions of whether or not the electron has been absorbed by the proton. I don’t mean that that isn’t a problem with a good (known?) solution, just I don’t know one.
- jiggawatts 6y agoPrecisely. This is the "measurement problem", and nobody can give you a straight answer. This is why I give this acid test: it reveals this gap in the theory that is otherwise easy to gloss over. It's the difference between pseudocode on a whiteboard, and something that compiles, runs, and gives the correct answer. You can hand-wave one of those and claim correctness. The other will give you a compiler error.
- drdeca 6y agoNo, I think this is just because you either haven’t been precise enough in what you want, or you are asking for something unreasonable. (Possibly a combination of the two) This is why I asked you to specify what you were asking for.
- forgotpwd16 6y agoSimulations of particle interactions play a significant role for physicists/engineers working in accelerators. See: https://www.fnal.gov/pub/today/archive/archive_2014/today14-10-21.html https://www.fnal.gov/pub/today/archive/archive_2014/today14-.... Indeed I also see a preference for symbolic simulations but numerical solutions are important and acceptable in all branches.
- jiggawatts 6y agoThat's a good start, but still fundamentally a cartoon. There's no visualisation of what goes on at the interaction point. Just an idealisation of what goes on far from it. Where in that diagram can I see diffraction? Interference? Entanglement? The particles interacting? Or can I not see any of that because it's a cartoon representation that shows the classical measurements only?
- forgotpwd16 6y agoNot sure what you mean with cartoon representation. It's classical measurement (position and velocity) which is what those building the accelerators need (trajectories and energies). This is very similar to satellite orbits simulations. But at the scale of interactions reality is dominated by probabilities. Each particle is basically a wavepacket. For an interactive animation of two colliding particles check: https://physics.weber.edu/schroeder/software/CollidingPackets.html https://physics.weber.edu/schroeder/software/CollidingPacket.... For an interactive animation of diffraction and interference check: https://physics.weber.edu/schroeder/software/QuantumScattering2D.html https://physics.weber.edu/schroeder/software/QuantumScatteri....
- rcxdude 6y ago> I've looked into why there's a total vacuum of numerical simulation, and the reason is simply snobbery. There's a philosophy that numerical methods are what mere engineers do, and not as intellectually pure as symbolic solutions. This is basically false as far as I can see. Numerical methods are basically the only way to solve either general relativity or quantum equations, but even then it's extremely difficult to get accurate calculations with a practical amount of compute power. There's intense competition over who can get the best results here (in fact the leading group in lattice QCD keeps their code a secret, which is not great science). I think the core reason you don't see much visualisations is the visualisations are not very useful. The little visualisations I've seen of extremely simplified versions quickly dissolve into complete mush as everything overlaps. It would not suprise me if the more complex interactions (like two real particles in 3d) were basically unintelligable and extremely difficult to make intelligable without becoming 'cartoons'.