I like the Lie algebra suggestion, but I doubt that, out of the box, it would support the tiny probabilities like 1e-100000 that Benjamin supports.

On Tuesday, April 4, 2023 at 11:51:11 AM EDT, Will Yager <will.yager@gmail.com> wrote:


If you end up coming up with a nice library design, you may want to generalize these operations from Double-valued logarithms to conversions between a Lie group and its corresponding Lie algebra.

Exp/ln on all/positive real numbers is a special case of exp/ln on a lie algebra/group.

Maybe worth checking if there already exists a package for working with lie groups in a nicely-typed way. If there's not a haskell one, I bet one of the theorem prover languages (coq, agda, idris) has one.

Best,
Will

_______________________________________________
Haskell-Cafe mailing list
To (un)subscribe, modify options or view archives go to:
http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
Only members subscribed via the mailman list are allowed to post.