[Octave-bug-tracker] [bug #30114] perms generates different results to matlab

[Qianqian Fang]

Follow-up Comment #5, bug #30114 (project octave):

On 06/11/2010 05:00 PM, Rik wrote:
> Follow-up Comment #3, bug #30114 (project octave):
>
> The documentation from Matlab
> (http://www.mathworks.com/access/helpdesk/help/techdoc/ref/perms.html)
doesn't
> guarantee any order to the permutations returned.  It seems that your
script
> should not be relying on any order since Matlab itself might change the
order
> between versions.
>    

I tried all versions of matlab (6~7.8), and the output orders
seem to be pretty consistent and predictable. In fact,
the help page does not say the orders are arbitrary either.

> I think it makes even less sense for Octave to try and match Matlab
behavior
> here since it is potentially a moving target.  Does this make sense to
you?
> Can you recode your script to not depend on the order of the permutations?
>    

I used permutation functions to generate tetrahedral
meshes from a grid. Because the orders of the elements are
not fixed, the indices of the element that encloses a 3D point
can vary. I could sort the elements, but that will add more
computation. I would rather rewrite the perms function.
If no one want to work on this, I will give it a try when I
have time.

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?30114>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/