| Re: bug in contour? [message #27355 is a reply to message #27280] |
Thu, 18 October 2001 01:06  |
Jacques
Messages: 5
Registered: September 1994
|
Junior Member |
|
|
Craig Markwardt wrote:
> Jacques <jfb37@NOSPAM.cam.ac.uk> writes:
>
>
>> Hi all
>>
>> Has anyone else run into a problem with the zlog keyword in contour? Or
>> can someone tell me where I am going wrong? It seems that if the keyword
>> is set, then any further contour plots are plotted using a log scale
>> irrespective of what value is sent to zlog. A way around it is to set
>> !z.type to 0 after each call to contour, but that shouldn't be the way
>> to do it.
>>
>> There is some simple code below to illustrate the point.
>>
> ... removed code ...
>
> I am not sure this is a bug. It is similar to the behavior of other
> /XLOG, /YLOG keywords, in the sense that once they are set, then any
> overplots are also performed in logarithmic coordinates. Since XLOG
> and YLOG are "sticky," I could see why ZLOG might be too. I will
They are only "sticky" within the same plot, which is good (even
necessary). But when a new plot is started, and the keyword is not set
then it should resort to the default, not inherit the properties from
the previous plot. Even worse in this case is that it is impossible to
override this inheritance - even if you explicitly tell it not to use a
log scale (zlog=0). You need to add the !z.type=0 line (or do a surface
plot inbetween or some other similar fix).
> grant that the ZLOG behavior is a bit surprising at first approach.
>
> The simple solution is to compute the logarithm yourself. I.e.,
>
> contour, ALOG10(dist(50))
Unless you want the original values on the contours (an aesthetic thing
in my case...)
> Craig
>
>
Jacques
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
