[Python-Dev] Rich comparison confusion

[Guido van Rossum]

>  > ... I can see no valid reason why someone would want to overload
>  > a>b different than b<a, ... 
>  > 
> 
> I agree.  But this assumes that the result of A<B and B>A is a
> collection of Booleans.  In the Interactive Data Language (IDL) these
> operators are essentially mapped to ceiling and floor functions which
> are not commutative.  I personally find this silly, but IDL users
> coming to Python may be surprised when the comparison of two Numeric
> arrays returns a Boolean-like result.

This means that Python can't be used to emulate this part of IDL.  I
don't understand how these can be not commutative unless they have a
side effect on the left argument, and that's not possible in Python
anyway.

--Guido van Rossum (home page: http://www.python.org/~guido/)