Re: [Gnuplot-bugs] problem .ps output with atanh(x)

[<HBB...@t-...>]

Erik Schaff. wrote:

> may be a bug?

No.  You just don't understand about sampling and finite resolution.

You're plotting a function that only exists in for -1 < x < 1, but over 
a range from -10 to 10.  Which means that you're wasting 90 percent of 
sampling points on nonexistent values.  The remaining 10 points at which 
the function actually returns a usable value come nowhere close enough 
to +/- 1.0 to generate the huge y amplitude you expect.

> I've just added the .ps output that I got with the expression atanh(x). Why does
> the plot stop at once ( around 2 on the y-axe), artanh has to be printed until +
> / - infinity as it reaches + - 1 on x-axe.

And how do you suppose infinity would fit onto a finite range of display?

I suggest you run an experiment: find the biggest number that's smaller 
than 1.0, and compute atanh() of that.  Explain why the result is so 
much smaller than infinity.