This would probably require a breaking change because tcspace() currently uses Stoddard and Prum's convention for vertices.

If we're in the mood of making breaking changes, we might as well go all the way through and use this black magic snippet from stack overflow:

simplex <- function(n) {
    qr.Q(qr(matrix(1,nrow=n)),complete=T)[,-1]
}

Woah, that's clever.

I'm open to a breaking change there if it means getting the gamut/monochromatic loci working. Using S&P has bugged me for a while anyway since it's out of line with the rest, but haven't had a good reason to swap. It's not a 'big' breaking change I guess since it's just a transform, so I'm probably on board if it means cool new functionality.

http://www.avidsynergy.com/pdf/n_dimensional_spherical_cordinates.pdf

This might be the last missing piece.

This would probably require a breaking change because tcspace() currently uses Stoddard and Prum's convention for vertices.

BTW, on that note, the vignette describes Kelber's convention, even for tetrachromats, when it's not the formula pavo currently uses.

Yeah, whoops. Let's consider the change for 2.4.

For safekeeping:

https://blog.r-hub.io/2020/02/10/code-generation/

This might be another option instead of a single function for all dimensions.