[Python-Dev] range objects in 3.x

[Stephen J. Turnbull]

Ethan Furman writes:

 > Well, actually, I'd be using it with dates.  ;)

Why are you representing dates with floats?

(That's a rhetorical question, don't answer it.)

This is the whole problem with this discussion.  Guido is saying (and
I think it's plausible though I don't have enough experience to  be
sure myself) that if you look at the various use cases for such
functions, they're different enough that it's going to be hard to come
up with a single API that is good, let alone optimal, for them all.
Then people keep coming back with "but look at X, where this API is
clearly very useful", for values of X restricted to "stuff they do".

That's good module design; it's not a good idea for the language
(including builtins).  Remember, something like range (Python 3) or
range (Python 2) was *really necessary*[1] to express in Python the
same algorithm that the C construct 'for' does.  I agree with Steven
d' that count would have been a somewhat better name (at least in my
dialect it is possible, though somewhat unusual, to say "count up from
10 to 20 by 3s"), but that doesn't become clear until you want to talk
about polymorphic versions of the concept.  Also, in statistics
"range" refers to a much smaller set (ie, {min, max}) than it does in
Python, not that I really care.<wink/>

As far as a name for a more general concept, perhaps "interval" would
be an interesting choice (although in analysis it has a connotation of
continuity that would be inappropriate for a discrete set of floats).

Footnotes: 
[1]  FSVO "necessary" that includes "let's not do arithmetic on the
index variable inside the loop".