Subject: hgfred and generalized hypergeometric functions
From: Raymond Toy
Date: Sun, 16 Apr 2006 09:45:59 -0400
Edmond Orignac wrote:
> I am using Maxima 5.9.1 on Debian Linux 3.1. I have recently discovered
> possible bugs in the hgfred function when applied to generalized
> hypergeometric functions such as 3F2 function.
> More specifically, I have compared the output of maxima with the
> identities listed on functions.wolfram.com for the generalized
> hypergeometric functions.
>
>
Can you provide a link to some of these identities?
> Generally, when the expressions reduce to a rational fraction,
> maxima gives the correct denominator, but the numerator differs from
> the formulas quoted in functions.wolfram.com except for the constant
> term.
>
>
There have been a large number of fixes to hgfred since 5.9.1. Please
try the 5.9.3 release or CVS.
If you really have 1000 or so tests, with the correct answers, I
certainly would like to have them. It's easy to make mistakes in
hgfred, so more tests are good.
> 5 5 1 1
> (%i3) HGFRED([-, -, 3], [-, -], x)
> 2 2 2 2
>
Currently maxima returns -(75*x^4+700*x^3+938*x^2+204*x+3)/(3*(x-1)^7).
Is that right?
Ray