[scilab-Users] min/max with NaN values

[Frederic Jourdin]

There are also inconsistencies even without NaNs involved.

Scilab/min is Scilab specific in the sense that one can run 
min(x1,x2,x3) in Scilab.
(it does not work in Matlab).
But Scilab/nanmin seems to be Matlab specific ( like min(x1,x2) in Matlab).

And the result is that Scilab/min and Scilab/nanmin are not
consistent each other.

For instance let's run in Scilab :

-->min(1,2)
 ans  =
 
    1. 
 
-->min(1,2,3)
 ans  =
 
    1. 
 
-->nanmin(1,2)
 ans  =
 
    1. 
 
-->nanmin(1,2,3)
              !--error 58
Wrong number of input arguments:

Arguments are :

 x         orient