Handling branch cuts for hypergeometric functions



A&S chapter 8 on Legendre functions is confusing. The same
name is used for different functions--only the name of the
independent variable (x or z) is changed. (Look at the
two formulae in 8.6.6.) I don't trust A&S 15.4.7---15.4.26.

Have you considered leaving 2F1 functions as 2F1 functions and
not using any one of 15.4.7---15.4.26 as simplifications?

Barton