Re: [Gnuplot-info] extrapolation beyond input range?

[Hans-Bernhard B. <br...@ph...>]

Dan Jacobson wrote:

> OK, I guess I wanted a way to get gnuplot to 'keep drawing those 
> great lines, don't chicken out when you come to the edge', but at the
> same time "don't ask me to derive some function to fit my data, just
> use the points I gave you".

Doing so would be spectactularly irresponsible.  Even human experts
don't really know how to do that job correctly, without background
information about the context this task comes up in.  And you want a
general-purpose computer program to just walk blindly ahead and do it?
No way.

Gnuplot calls itself a _scientific_ plotting program.  What you're
requesting would be quite exactly the opposite of "scientific", and we
won't do it.

> I mean can't gnuplot make some best fit parabola or something for the
> given points?

Sure.  But there's no way it could ever make the decision between that
parabola and "or something" for you.