| PV-Wave plotting adventures [message #15658] |
Wed, 02 June 1999 00:00 |
luthi
Messages: 20
Registered: March 1999
|
Junior Member |
|
|
Here is a little caveat for all users of PV-Wave plotting routines (I hope
this is fixed in IDL!). The errors appeard in PV-WAVE v6.10 (Solaris) but as
far as I remember were in all earlier versions (however I did not fully
realize the reason of this sometimes strange behaviour...)
PV-Wave plotting adventures
===========================
A really annoying plotting bug appears when the range of the plotting region
is some orders of magnitude less than the frame offset:
x0=1.0 ; frame offset
dx=1.0d-5 ; plotting range
plot,[0],[0],xrange=[x0,x0+dx],yrange=[0,1],/xstyle,/ystyle
The plot is either inaccurate, unwanted tick marks appear inside or outside
of the plotting region and the axes labeling gets somewhat random (switching
between decimal and exponential notation)
Try for example
x0=1.0
dx=1.0d-4
plot,[0],[0],xrange=[x0,x0+dx],yrange=[0,1],/xstyle,/ystyle
and
x0=1.0
dx=3.0d-4
plot,[0],[0],xrange=[x0,x0+dx],yrange=[0,1],/xstyle,/ystyle
Unusable annotations and tickmarks outside of the plotting frame can be
produced with
x0=10.0
dx=1.0d-5
plot,[0],[0],xrange=[x0,x0+dx],yrange=[x0,x0+dx],/xstyle,/ys tyle
getting worse and worse as x0 increases (even if the plotting intervals are
always the same size). A fully collapsed plot is produced with
x0=1000.0d0
dx=1.0d-5
plot,[0],[0],xrange=[x0,x0+dx],yrange=[x0,x0+dx],/xstyle,/ys tyle
also using double precision numbers. Increasing dx yields this ugly plot where
the tick marks on the vertical axis are NOT EQUALLY SPACED!!!!
x0=1000.0d0
dx=1.0d-3
plot,[0],[0],xrange=[x0,x0+dx],yrange=[x0,x0+dx],/xstyle,/ys tyle
Using plot_dbl gives the same crazy results, unless one uses the keyword bits
xstyle=32, ystyle=32
x0=1000.0d0
dx=1.0d-3
plot_dbl,[0.0d0],[0.0d0],$
xrange=double([x0,x0+dx]),yrange=double([x0,x0+dx]),xstyle=3 3,ystyle=33
To be sure I display the header of plot_dbl here:
;$Id: plot_dbl.pro,v 1.1 1995/09/18 15:19:37 landers Exp $
; Copyright 1994 Visual Numerics, Inc.
This means that the plotting bugs described above have been known by VNI for
at least four years and at least two major releases without fixing them!
Using plot_dbl solves the problem of labeling the axis only partly, since one
has to use oplot_dbl (which also works fine) but where are all the other
plotting routines like contour_dbl, surface_dbl aso.?
And having all those routines, why not simply fixing this annoying bug in the
main plotting routine.
I guess that many of the date-time plotting errors are really caused by the
behaviour of the plotting routine when the plotting range is some orders of
magnitude smaller than the offset (which is almost always the case since
plotting date/time uses julian date for plotting).
PLEASE any people at VNI reading this, PLEASE FIX THIS BUG IMMEDIATELY!!!
Cheers Martin
--
============================================================
Martin Luethi Tel. +41 1 632 40 92
Glaciology Section Fax. +41 1 632 11 92
VAW ETH Zuerich
CH-8092 Zuerich mail luthi@vaw.baum.ethz.ch
Switzerland
============================================================
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
