>>>>> "Barton" == Barton Willis writes:
Barton> (1) What is the (A & S or G & R) name of the %P function in
Barton> (%i5) hgfred([a+1/2,a],[c],x);
Barton> (%o5)
Barton> (2^(1-c)*%P[2*a-c,1-c](1/sqrt(1-x))*(1-x)^(c-2*a-1/2)*x^(-2*c+2*a+1))/(g
Barton> amma(c)*(x^2-1)^((1-c)/2))
Hmm. You should update your copy. %p has been replaced by
assoc_legendre_p.
Barton> (2) Where can I find a reference to the identity %i5 = %o5 ?
Unfortunately, you'll have to read the code. I've tried to leave
comments on what formula is used or derivations used. Mostly
references to A&S.
Finding out what formula is used isn't easy. Usually you have to
trace the code, or put in formats. However, for %f[2,1] functions,
you can set trac2f1 to true, and watch the various tests to see how it
arrives at this answer.
Barton> (3) Does the message 'simp2f1-will-continue-in' have any meaning
Barton> that a user should know about? For example
Barton> (%i6) hgfred([a+1/3,a],[c],x);
Barton> SIMP2F1-WILL-CONTINUE-IN
Barton> (%o6) %F[2,1]([a+1/3,a],[c],x)
Not really sure, but I guess the author thought there are other cases
where we should be able to simplify %f[2,1] functions, but we haven't
implemented them.
I've left most of these in but they should eventually go away if we
ever figure out what they really mean or if we solve the issue.
Ray