[Numpy-discussion] (Late) summary of PEP-225 discussion at Scipy

Robert Kern robert.kern@gmail....
Wed Oct 29 14:43:45 CDT 2008

On Wed, Oct 29, 2008 at 14:33, Eric Firing <efiring@hawaii.edu> wrote:
> David Warde-Farley wrote:
>> Hi Fernando,
>> In Robert's comment, I think the reST processor somehow got rid of a
>> backslash. In my browser I see
>>       (I'm looking at you, Matlab's "")
>> although this is an aside that will be lost on anyone who hasn't used
>> the Matlab backslash operator anyway.
>> In fact, one other thing just came to mind that isn't obvious and (so
>> far as I can remember) hasn't been brought up: even though Robert and
>> others (myself included) really only care about an operator to
>> facilitate matrix multiplication, I think most of us would support the
>> ~*=  (mentioned in the PEP) as a natural companion to ~*.
> This makes no sense to me at all.  Why use in-place notation for an
> operation that is not in-place?  The point of having a matrix
> multiplication operator is to facilitate code that corresponds more
> closely to mathematical notation.  The in-place operations live in the
> computer science world, not the mathematical world.  The proposal for
> ~*= violates language consistency while doing nothing to improve
> mathematical readability.

Eh, that's not entirely true.

  x = 1
  x += 2

That's not in-place. They are called "augmented assignments", not
"in-place operations" for this reason. The defining characteristic is
that "x <op>= y" should be equivalent to "x = x <op> y" except
possibly for *optional* in-place semantics.

But this isn't a choice for us to make. If we add operators to the
Python language, their associated augmented assignments will also be
added (no matter what we think about it). If a class doesn't define
the appropriate __magic__ method, then Python will execute "x = x <op>
y" and give us the semantics that David was suggesting. This is a

Robert Kern

"I have come to believe that the whole world is an enigma, a harmless
enigma that is made terrible by our own mad attempt to interpret it as
though it had an underlying truth."
  -- Umberto Eco

More information about the Numpy-discussion mailing list