A rotating triangular parabolic resonator

Like the simulation this one shows a rotating parabolic resonator, this time with 3 sides instead of 8. Parabolas have the property of turning a circular wave coming from their focus into a linear wave, and vice versa. For a static resonator, as in the video , the linear wave created after the first reflection will not hit the opposite sides at the correct angle to be transformed back into a circular wave. However here, the speed of rotation is such that the linear waves actually do hit another side at approximately the right angle, allowing to recover a nearly periodic pattern. The slow dispersion as time goes on may be partly due to round-off errors near the parabolic boundaries, and appears to be increased by the rotating boundary conditions are implemented in the simplest possible way by setting the wave to zero outside the rotating octagon. They thus do not take into account any energy that the boundary might transfer to the wave. T