Rotation matrix may be computed just once

This might speed up the simulation, but probably not enough to have an acceptable frame rate in real time. So I guess you would still need to make a time lapse. In that regard, I'm not sure this is worth the trouble.

What the code does now is increment the rotation angle every loop by 1 degree, construct a new rotation matrix and use that matrix only once. Alternatively, you could construct a rotation matrix just once, before the draw loop, with an angle of 1 degree. Then use that matrix every time to transform the current coordinates. This will have the same effect, cumulatively. So instead of using sin and cos every loop *and* do the matrix multiplication, you have only the multiplication left.

One slight complication is that matrix multiplication p' = R.p can't be done in-place, because the old values would be overwritten when you still need them. So in BASIC, this probably requires juggling with array indices where you transform p(0..7) into p(8..15) on one loop, and p(8..15) into p(0..7) on the next loop.

I hope this is clear and that it will be helpful. Cheers!

Start at the draw loop and trace how the rotation angle, rotation matrix, and coordinate arrays p(0..7) and p(8..15) are used. Compare the current per-loop matrix construction with reusing a fixed 1-degree matrix and alternating the coordinate arrays; done means the cumulative rotation has the same effect without rebuilding the matrix each loop.

Scritto dal modello di indicizzazione a partire dal testo della issue.