I believe the problem is not about negative strides but about negative bounds. There should be a notion of "minus zero", something like

"abcde"[:-0:-1] =="edcba".

Here ":-" serves as a special syntax for negative stride; of course it is not a real proposal.
The same awkwardness results when you take a negative upper bounds to the limit of 0:

"abcde"[:-2] == "abc"
"abcde"[:-1] == "abcd"
"abcde"[:-0] == "" 

(I once filed a bug for it, which was of course correctly rejected: http://bugs.python.org/issue17287).


2013/10/27 Guido van Rossum <guido@python.org>
In the comments of http://python-history.blogspot.com/2013/10/why-python-uses-0-based-indexing.html there were some complaints about the interpretation of the bounds for negative strides, and I have to admin it feels wrong. Where did we go wrong? For example,

"abcde"[::-1] == "edcba"

as you'd expect, but there is no number you can put as the second bound to get the same result:

"abcde"[:1:-1] == "edc"
"abcde"[:0:-1] == "edcb"


"abcde":-1:-1] == ""

I'm guessing it all comes from the semantics I assigned to negative stride for range() long ago, unthinkingly combined with the rules for negative indices.

Are we stuck with this forever? If we want to fix this in Python 4 we'd have to start deprecating negative stride with non-empty lower/upper bounds now. And we'd have to start deprecating negative step for range() altogether, recommending reversed(range(lower, upper)) instead.

Thoughts? Is NumPy also affected?

--Guido van Rossum (python.org/~guido)

Python-ideas mailing list