So there is a bit of confusing usage of squareness. sqfactor vs tau.

tau is the linearized parameter for squareness, and tau is used in the papers equations,
The nonlinear "s" parameter (say sqvalue), of squareness is a function of tau.
The use of tau as a label for node input, does not mean anything to a user.

sqfactor seems like a good label for the node input parameters. sqfactor = 1 produces a square, (with slightly rounded corners.)
So sqfactor = tau is good to use.
("squareness" seems to be a good label, except for its length.)

A different name from sqfactor should be used in the function eval_rho. say sqvalue.
var sqfactor = (function of tau); should be replaced with something like:
var sqvalue = (function of tau);

- just an opinion,
- Brian