Handling branch cuts for hypergeometric functions

>>>>> "Barton" == Barton Willis  writes:

    Barton> The x vs z scheme works OK in a reference book, but it doesn't
    Barton> work for a CAS.  Either Maxima needs different names for the
    Barton> two definitions of the associated Legendre functions (A&S 8.3.1, 8.1.2),
    Barton> or Maxima needs to stick to one or the other. The difference
    Barton> between these two definitions cannot be explained by saying
    Barton> that the factor ((z + 1)/(z-1))^(mu / 2) is multivalued. The
    Barton> functions are different.

I'm open to suggestions.  And, in fact, I'm using your orthogonal
polynomials to deal with this, so I can do whatever you do.

Ray