Re: NLEVELS doesn't necessarily mean NLEVELS

[Struan Gray <struan.gray(at)sljus.lu.se>]

David Fanning, davidf@dfanning.com writes:

> I contacted RSI technical support and learned that contrary
> to what the IDL documentation claims, the NLEVELS keyword
> does not divide the data range into "this number of equally
> spaced contour intervals". Rather, this is taken as a
> (I'm not kidding) "suggestion" to the CONTOUR command, and
> the CONTOUR command can apparently do what it likes with the
> suggestion. Including, as I understand it, put it where the 
> sun don't shine.

    When you're tired of this one, try playing with the BINSIZE keyword
in calls to HISTOGRAM and see if you can predict how many bins your
histogram will have.  I never could.


Struan