I wonder if it would not be easier to simply write down equations defining the bisector in terms of the various parameters, and see if it's solvable with square roots.
I wonder if it would not be easier to simply write down equations defining the bisector in terms of the various parameters, and see if it's solvable with square roots.