When running a simulation with the rectangle structure, there's an extra reflection in the bottom-left corner. Doesn't seem to do anything on initial field generation, but when the field reflects off of the block it shows up.

It appears (and I believe it should be) that every point in the stored fields for the empty run are the same for a given frequency (i.e. Ey(0,yPos) == const for all yPos). Re-using the same data point for each position will save memory and make the stored files significantly more legible.

Currently emtpy_refl_data.h and empty_tran_data.h are overwritten on every call of ./empty, but they do not automatically write in the text for the license at the top.

As it stands coming up with material fits is at best sort of confusing. It uses both Matlab and Mathematica and can still produce fits that cause the simulation to become unstable.

• Update to include Levy estimation of initial parameters in Mathematica (or do the advanced fitting in Matlab, either way works).

• Try to establish criteria that ensure stability of a material model in the simulation. This might be an entire PhD though, so don't get too caught up on it.

Not sure yet the cause, but the DFT and/or Reflection and Transmission results are super weird

Output can be hard to interpret when R -> 0 and T -> 1 due to the scale, so it would be handy to scale the chart based on the max value of each dataset.

Update the data download script to include relevant information for what was simulated:

• Background Environment
• Shape
• Size
• Adaptive (yes/no)
• Material

Right now it seems like all the metals are basically the same (in terms of results). There are slight differences, but by and large they aren't particularly meaningful. Adjust the modelling to use Lorentz-Drude model instead of just Drude approximation to try and get more physical results.

Comment received that being more clear about what X and Y mean (as size dimensions) would be useful. Add little diagrams that update depending on which shape is selected.

The outline of the objects is off by a singe pixel (I think)

Currently number of steps in the simulation is hard-coded -- now that users can input high dielectric values for the object it is possible that the number of steps will not be enough for the energy to sufficiently dissipate (particularly when on-resonance). Can use ArrayMax to find current field strength and wait for it to decay to a sufficiently low value before terminating in these conditions.

• Add figure heading and axes labels
• Add checkbox so users can store data from previous runs to overlay with new data (this will require auto-renaming the Chart.js datasets so they can be distinguished properly...)
• Option to download datasets shown on the chart (and / or raw e-field strength as a function of time?)

Sometime last night the error:
Looks like you are rendering without using requestAnimationFrame for the main loop. You should use 0 for the frame rate in emscripten_set_main_loop in order to use requestAnimationFrame, as that can greatly improve your frame rates!

started showing up. It didn't use to, so it must be something that I changed somewhere along the way. Not sure if it's technically an issue as everything still works, but I'm not happy that it's so unhappy.