[Octave-bug-tracker] [bug #65238] Improve nchoosek.m algorithm to prevent numerical issues

[Markus Mützel]

Update of bug#65238 (group octave):

                  Status:                    None => Need Info              
        Operating System:               GNU/Linux => Any                    
         Planned Release:                    None => 10.1.0 (current default)

    _______________________________________________________

Follow-up Comment #1:

Thank you for your contribution. The proposed changes look good to me. It's
not often that a change improves both the numberical stability *and* the
performance. 👍

I made some minor changes (mostly style, see the attached patch):
* Don't use CamelCase for variable names.
* Use space between function name and opening parenthesis.
* Remove trailing whitespace.
* Use space after negation operator "!".
* Use "##" for comments that start at the beginning of a line.
* Minor rewording of comment block.
* Add a commit message.

Some questions:
* Is there a reason why the multiplication and division is done in two steps
for the "default case" (lines 145-146), but it is done in one step for the
"large number case" (line 138)?
* Is your name and email in the patch still correct?
* Does the commit message correctly describe the change?
* Could I add you to the list of contributors that appears at the beginning of
the manual? https://docs.octave.org/latest/Acknowledgements.html


(file #55660)

    _______________________________________________________

Additional Item Attachment:

File name: bug65238-nchoosek-v1.patch     Size:3 KB
   
<https://file.savannah.gnu.org/file/bug65238-nchoosek-v1.patch?file_id=55660>


    AGPL NOTICE

These attachments are served by Savane. You can download the corresponding
source code of Savane at
https://git.savannah.nongnu.org/cgit/administration/savane.git/snapshot/savane-7984227a52b1c59ed77ba5c9f7ef631497509617.tar.gz


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?65238>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/