rtest14 and specint results



Raymond Toy wrote:

>>>>>>"Richard" == Richard Fateman  writes:
> 
> 
>     Richard> For what it's worth, commercial macsyma seems to fail on
>     Richard> these problems with Proviso: Assuming p>0
> 
> Bummer.  I was hoping for independent verification.  Anyone with
> Mathematica who can try out those integrals?

Mathematica...
InputForm[HypergeometricPFQ[{1, 2}, {3/2}, z]] =

(-(Sqrt[1 - z]*Sqrt[z]) - ArcSin[Sqrt[z]])/
  (2*Sqrt[1 - z]*(-1 + z)*Sqrt[z])

with a -z  in there, it produces
(Sqrt[z]*Sqrt[1 + z] + ArcSinh[Sqrt[z]])/
  (2*Sqrt[z]*(1 + z)^(3/2))

What exactly would you like to see done in Mathematica?  There is
a web site
http://integrals.wolfram.com/
that will do indefinite integrals.
RJF

> 
> Can macsyma simplify hypergeometric functions?  For example,
> expressing F(1,2;3/2;z) in closed form?  Maxima produces 2 answers
> depending on whether z is positive or negative.  I've been doing that
> by hand in rtesthyp.mac, and it's quite tedious and error-prone,
> especially when the resulting answer is long.  An independent
> verification would be good.
> 
> Ray
> 
> 
> _______________________________________________
> Maxima mailing list
> Maxima@www.math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima