element-wise division is more important than floats
Reported by adrien vogt-schilb
Originally assigned to Pierre-Aime AGNEL
Hi
there is a classics in scilab:
M= [1 2; 3 4];
1./M
gives the matrix inverse of M, because the parser gets :
(1.)/M, ie 1.0/M or 1/M
in my short experience, everybody has a piece of code designed in order to get the element-wise inverse of a matrix, "wrongly" coded in a 1./M fashion.
This "bug" is very well known by specialists, they know that it can be easily worked around :
//getting the element-wise inverse of M. Do not remove the space before the point.
1 ./M //note the space
but still, who would want 1./M to be the matrix inverse of M?
I suggest that in scilab 6.0, the new parser should understand 1./M as element-wise inverse.
I am aware of course that this could turn out to be extremely difficult to implement for some reason, but i think it is worth asking.