Check if a given value is out of certain range
Steven D'Aprano
steve at pearwood.info
Sat Oct 3 06:29:05 EDT 2015
On Fri, 2 Oct 2015 07:45 am, John Gordon wrote:
> In <87r3le1ht3.fsf at elektro.pacujo.net> Marko Rauhamaa <marko at pacujo.net>
> writes:
[...]
>> Wouldn't
>
>> x < 0 or 10 < x
>
>> be even more visual?
>
> I don't know what you mean by "more visual".
>
> In my opinion, when comparing a variable to a constant, it's more natural
> to have the variable on the left and the constant on the right, so that's
> one strike against this code.
>
> Another strike is that the code isn't consistent with itself; it puts the
> variable on the left in the first comparison, then swaps to the right for
> the second comparison.
The alternative is equally inconsistent: for one comparison it uses less
than, for the other it uses greater than.
I find this discussion about the relative readability of
not 0 <= x <= 10 #1
versus
0 < x or x > 10 #2
0 < x or 10 < x
to be a good example of people's propensity to invent so-called "rational"
justifications for irrational preferences. Honestly, all these people
claiming that reading #1 takes them "slightly more effort" than reading #2,
or vice versa. Really? You've done an objective test of this? I don't think
so.
We know from UI testing in other fields that people's *self-reported*
efficiency and their *actual* efficiency are almost completely
uncorrelated. Back in the DOS versus Macintosh days, when mice were new,
DOS users consistently reported that using a mouse was too slow and
inefficient compared to the keyboard, but actual objective measurements of
the time that they took to do tasks showed the opposite.
E.g. subjects were asked to do some task, using the keyboard and then again
using the mouse, and (let's say) it actually took them 3 minutes to do it
via keyboard and 2 minutes via mouse, they reported that using the keyboard
commands was much more efficient, fast and easy.
Whether you have #1 or either variation of #2, the time it takes to read and
comprehend the expression is likely to be of the order of a few dozen
milliseconds, plus or minus a few dozen milliseconds. I expect the
variation will be nearly as great as the average time. (But of course, I
haven't done objective studies either.)
Anyone who has done high-school level maths should be familiar with the
notation `0 <= x <= 10`, read as "x between 0 and 10". (Anyone who is not
can hardly claim to be fluent in Python -- that's like claiming to be
fluent in English while not knowing where and how to use an exclamation
mark.)
If you want to check for the opposite, x not between 0 and 10, the natural
way to do so is with `not`: `not 0 <= x <= 10`. If you prefer `x < 0 or x >
10` for whatever reason, that's fine too. Some people prefer pizza, some
prefer kebabs... inventing (spurious) arguments of (imaginary) efficiency
("It takes me half a second less time to chew each mouthful of burger
compared to kebab") is unnecessary. But we do love the argument from
efficiency.
Frankly, as a profession, we programmers are lousy at prioritising. We write
pages and pages of convoluted, inconsistent, hard to understand code that
can sometimes take hours or even days to comprehend well enough to make
even a minor change, while spending dozens of person-hours arguing about
which expression saves them a microsecond or two of reading time. Yay us!
--
Steven
More information about the Python-list
mailing list