[Maxima-bugs] [ maxima-Bugs-2892329 ] Plot2d sometimes incorrect when using logscale

[SourceForge.net <no...@so...>]

Bugs item #2892329, was opened at 2009-11-05 02:16
Message generated for change (Settings changed) made by crategus
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=2892329&group_id=4933

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Lisp Core - Plotting
Group: None
>Status: Pending
>Resolution: Wont Fix
Priority: 5
Private: No
Submitted By: William Holtz (wmholtz)
Assigned to: Nobody/Anonymous (nobody)
Summary: Plot2d sometimes incorrect when using logscale

Initial Comment:
Using 5.19.2 on winXP.

plot2d([1/(x+1)], [x,1e-3,1e3], [logx])$
plot2d([1/(x+1)], [x,1e-3,1e3], [gnuplot_preamble, "set logscale x"])$

These should give identical plots, but I the second version gives me a "knee" in the plot that is incorrect. This only happens with certain x ranges. I have attached an image of the incorrect plot.

Please let me know if there is additional information I can provide -- great program overall!

----------------------------------------------------------------------

>Comment By: Dieter Kaiser (crategus)
Date: 2009-11-12 11:33

Message:
As described in the last posting it is expected that the plots are
different, because of a different sampling of points for the plot. I think
this is not a bug. This can be seen when using the style points to plot the
function:

plot2d([1/(x+1)], [x,1e-3,1e3], [style, points], [gnuplot_preamble, "set
logscale x"])$
plot2d([1/(x+1)], [x,1e-3,1e3], [style,points], [logx])$

Setting this bug report to pending and the resolution to "Wont Fix".

Dieter Kaiser

----------------------------------------------------------------------

Comment By: Raymond Toy (rtoy)
Date: 2009-11-05 02:31

Message:
The issue is how maxima does plots.  Basically, plots are done by uniformly
sampling the x range.  (Roughly).  But when you want to do a log plot, the
x range should be sampled logarithmically.  When you specify logx, maxima
knows to do the sampling correctly.  That's why you get a nice smooth
curve.

But with the preamble, maxima doesn't know you're producing a log plot, so
it does the normal sampling.  That's why you see the knee.  The samples
selected are at 0.001 and about 0.78.

(The sampling is roughly uniform, because maxima has an adaptive algorithm
that will sample more if the plot is not smooth enough.)



----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=104933&aid=2892329&group_id=4933