[Gnuplot-info] question for extension definition range

[Francky L. <Fr...@Le...>]

Dear,

I have a problem with ranges.
Attached is my example.

The blue curve (error) is an oscillation function between the error lines
at 1.67 and -1.67.

samples=101.eps:
obtained with set samples 101
As you can see the oscillating function practically
never reaches the limits. This is not because the function is wrong,
but because the sampling is insufficient. Therefore we increase the number
of sampling points.

samples=1001.eps:
obtained with set samples 1001
Much better, but still not sufficient.
Some of the extrema are not reached.

The problem is that the sampling x doesn't necessary coincides
with the x value of an extremum.

Therefore, I propose the following extension of the range of the
plot command:

plot  [0:x1:x2:x3:x4:x5:x6:3.5] ln_ITcan_log(x) axes x1y1  notitle with 
lines  linetype 1  linewidth 5, \
                                             log_error_log(x) axes x1y2  
notitle with lines  linetype 3  linewidth 5

between each consecutive x values for the range a number of samples defined
by set samples would be taken. That way one would obtain:

1) a polyline wich is continuous.
2) a polyline wich reaches the extrema. Indeed, the x values of the
   extrema would be sampled.

What about the idea?

Best regards,

Francky Leyn