Barton,
Thanks for these great contributions.
Robert suggests:
> I wouldn't want to preserve nummod as a separate function
> apart from the existing function mod; we should modify mod
> so that it calls nummod for the appropriate cases.
Unfortunately, I don't see any consistent way to do this.
mod(x,2) => x
mod(n*(n+1),2) => n^2+n
but
nummod(x,2) == nummod(x,2) (noun form)
nummod(n*(n+1),2) => 0 (declare(n,integer))
It seems to me that nummod is more fundamental than the current mod,
which should be renamed polymod or some such, leaving the name mod for
nummod. How much chaos would this cause?
-s