Ricky Reusser

Least Squares Curve Fit with the QR Decomposition

Jacobi Elliptic Functions in the Complex Plane

Computing π with the Bailey-Borwein-Plouffe Formula

Adaptive Contouring in Fragment Shaders

Grouping Points with Principal Component Analysis

Domain Coloring with Adaptive Contouring

Double Compound Pendulums

3D Reaction-Diffusion

Quasicrystals

2D (Non-physical) N-body Gravity with Poisson's Equation

Periodic Planar Three-Body Orbits

eases

Finding Roots in the Complex Plane

glsl-complex

glsl-complex-auto-differentiation

polynomial-roots

complex-auto-differentiation

utils

Integration

Math

Here's a strange bug: adaptiveSimpson(x => x * Math.sin(n * x), 0, Math.PI) is effectively zero for all n that are multiples of 4, though the correct result is +/- pi/n. I discovered this while trying to develop some tools for Fourier Series. :)

Mark McClure

Interesting. Is it a case where it samples all of the zeros, perhaps? I could work in *minimum* recursion depth parameter.

I've added a minDepth parameter to help overcome this, though you've hit on a weakness of this method. I've also allowed use of a params object instead, so that, for example, `adaptiveSimpson(x => x * Math.sin(n * x), 0, Math.PI, { minDepth: 2 })` fixes this particular case. It's not ideal since there's not a clear indication when you've aligned the zeros with the function's zeros, but it's the best I have for now! Thanks for pointing out this case!

*when you've aligned the zeros with the sample points