Hi Patrick,
Actually I didn't think so; but upon examination it seems your
right about the infinite being a correction (of some sort). But
calling the infinite sum with these values is causing gamma to be
called with a zero; which it shouldn't be. Correcting the finite sum
with an infinite one is not what I expected.
Sigh.. I will look into what is going on in the infinite sum more
carefully. This is not the easiest code to follow :) This time I will
start altering the code to check my analysis, and write down the program
flow. (:
Ray
On 03/14/2014 02:29 PM, Patrick Alken wrote:
Hi Ray,
Thanks for your report. If you have a correction you can upload it
to the bug tracker or just post it to this list.
I don't know much about these functions but maybe someone else on the
list can check it. Based on that code snippet, it looks like it starts
off with a finite sum approximation to the true value, and then
refines that approximation with an infinite sum? It seems to pass in
the finite sum value to the infinite sum routine.
But you're saying under certain argument conditions, its not necessary
to refine the initial approximation with the infinite sum?
Patrick
On 03/14/2014 12:24 PM, Raymond Rogers wrote:
I have added an analysis to the bug report. Somebody needs to double
check it; I usually get things a little wrong the first time through.
If it should happen that my suggestion is valid; how and where do I
submit a correction?
Ray
https://savannah.gnu.org/bugs/index.php?41837#comment4
On 03/11/2014 06:24 PM, Patrick Alken wrote:
Additional Item Attachment, bug #41837 (project gsl):
File name: gsl_sf_hyperg_U_bug_report.cpp Size:0 KB
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/bugs/?41837>
_______________________________________________
Message sent via/by Savannah
http://savannah.gnu.org/
--
Act IV, Sc. IV
What is a man,
If his chief good and profit of his time
Be to sleep and feed. Be a beast, no more
