restriction on sum: intentional bug?

Jon Clements joncle at
Fri Oct 16 20:06:30 CEST 2009

On Oct 16, 5:59 pm, Tim Chase <python.l... at> wrote:
> Stephen Hansen wrote:
> >> Why doesn't duck typing apply to `sum`?
> > Because it would be so hideously slow and inefficient that it'd be way too
> > easy a way for people to program something they think should work fine but
> > really doesn't... alternatively, the function would have to do two
> > /completely/ different implementations based on what you're passing in, and
> > that violates duck typing too :)
> But that's the issue...supporting strings does NOT involve "two
> /completely/ different implementations based on what you're
> passing in" but rather just a reduction of starting point
> (whether 0 or '') and an object that had an __add__ method.
> String meet these requirements.  Specifically disqualifying
> strings is where you get the two code-paths/implementations.
> So I agree with Alan & Peter that this creates an unfortunate
> language wart among (as Peter aptly puts it) "consenting adults".
>   I'd feel similarly if certain classes anomalously prevented
> access to internal "private" data.  We know that if we go mucking
> around like this, it's our own fault if things break or get slow.
>   Is Python going to prevent me from typing
>    count = 0
>    for i in range(1000000):
>      if i % 1000: count += 1
> instead of specifying the step-size?  Or even forcing me to
> precompute this constant value for `count` because looping is
> inefficient in this case?
> Even more annoyingly, sum() *knows the right thing to do* to the
> degree that it even puts it in the error message...INSTEAD OF
> JUST FREAKIN' DOING THE RIGHT THING.  Do it inefficiently, or do
> it efficiently, but it's NOT AN ERROR. (IMHO ;-)
> </soapbox>
> -tkc

Does it know the right thing to do though?

For instance, sum(['1', '2', '3']); it's not completely unreasonable
for someone to expect
a result of 6.

No one seems to mind that ''.join([1,2,3]) baulks, and doesn't return

Explicit is better than implicit.


More information about the Python-list mailing list