Change the anchor @esperanc/masson-elliptic-lib to @esperanc/masson-non-trivial-math

Fix the link to Ricky Reusser's notebook: https://observablehq.com/@rreusser/periodic-planar-three-body-orbits

For a specific value of m, a rational approximation can be very fast. See https://observablehq.com/@jrus/fast-peirce-quincuncial (you also might enjoy https://observablehq.com/@jrus/scpie)

Jacob thanks for the suggestion. However take a look at Fukushima's approach: piecewise polynomials w pre-calculated coefficients.

I am thinking about computing the (complex-valued) *incomplete* integral for a specific value of *m* and its inverse (as in the code below in this notebook), including with good precision near the singular corners, where many of the existing approximations have some difficulty. Computing the complete integral (as in the papers above) is a slightly different problem (though I think a rational approximation can potentially be helpful there too). My notebook above is just the disk-to-square case; I thought about but didn’t ever get around to implementing a disk-to-arbitrary-rectangle code (and computing the coefficients in Javascript would take porting a bunch of other building blocks).

Awesome, please keep me posted. I left the incomplete integral as a to-do port task, Fukushima has F90 code for that as well.