[Numpy-discussion] Name change of the ptp() function

David Cournapeau cournape at gmail.com
Fri May 10 09:53:33 EDT 2013

On Fri, May 10, 2013 at 2:27 PM, Ralf Gommers <ralf.gommers at gmail.com> wrote:
> On Fri, May 10, 2013 at 3:17 PM, Robert Kern <robert.kern at gmail.com> wrote:
>> On Fri, May 10, 2013 at 2:04 PM, Eli Bressert <ebressert at gmail.com> wrote:
>> > That's a good point regarding the range function names. But, I think
>> > the issue still stands on the readability of the ptp function.
>> > Regarding PEP20 it's stated that "readability counts."
>> >
>> > If you regard what ptp is supposed to replace, array.max() -
>> > array.min(), the aforementioned follows the PEP20 better as it is more
>> > readable. If valuerange() is not an acceptable name, maybe span()?
>> Sure, it's probably more readable, and that would be the controlling
>> factor if this were a new function. But that's not really the
>> operative question here. Are the gains in readability worth the
>> nontrivial costs of deprecating and removing the old name? I, for one,
>> am generally not in favor of such deprecations.
> That's not the only option though. I'm -1 on deprecation, but +0 on renaming
> and keeping ptp as an alias. The function name is really quite poor.

I think it is a matter of context. I don't know the history of that
function, but coming from a signal processing background, its meaning
was obvious to me. It is a peak to peak is a very common operations
when dealing with audio file, for example (that's how most wave
display work AFAIK).

I am certainly -1 on the deprecation as well, and -0 on alias.


More information about the NumPy-Discussion mailing list