Re: Undocumented behaviour of "smooth" options

[Allin C. <cot...@wf...>]

On Thu, 23 Feb 2017, Hans-Bernhard Bröker wrote:

> Am 22.02.2017 um 00:21 schrieb Ethan A Merritt:
>
>> On the bug tracker    <https://sourceforge.net/p/gnuplot/bugs/1911/>
>> Rainald Koch points out that when a data set contains some points with
>> invalid y values, the result of smoothing is unexpected.
>
> Expectations can be deceiving.  :-)
>
>> I can't find any mention of it in the documentation, but it seems that
>> this is intentional.  The "smooth" code has a pre-processing step that
>> splits the input data into N subsets separated by one or more "undefined"
>> points.
>
> Well, what else was the code to do with actual blank lines in the input,
> or otherwise missing/unusable data?  And back in the day, invalid data
> _was_ often treated as being the same thing as missing data.  Plotting a
> spline across a hole in its input data would IMHO be at least as wrong
> as the lack of documentation of the current behaviour.
>
> The behaviour of the smoothing code is not that wildly different from
> that of the normal 'with lines' behaviour, really: a missing entry
> breaks the polyline, so it's not entirely unepected that it breaks the
> smoothed curve, too.

That seems reasonable. If the user has reason to believe the missing 
data can be interpolated in some way, she could do that before 
smoothing.

Allin Cottrell