RE: Strange Results for Resampled Irregular3DSet

• To: "'Curtis Rueden'" <ctrueden@xxxxxxxx>
• Subject: RE: Strange Results for Resampled Irregular3DSet
• From: "STAAB, Eugen" <eugen.staab@xxxxxxxxxxxxx>
• Date: Fri, 10 Feb 2006 08:55:47 +0100

Thanks for the help, Curtis!

Best wishes,
Eugen

-----Ursprüngliche Nachricht-----
Von: Curtis Rueden [mailto:ctrueden@xxxxxxxx]
Gesendet: Donnerstag, 9. Februar 2006 22:40
An: STAAB, Eugen
Cc: 'visad@xxxxxxxxxxxxxxxx'
Betreff: Re: Strange Results for Resampled Irregular3DSet


Hi Eugen,

I said:

> Fortunately, you may not need to use IrregularSets at all. If the data 
> is merely unevenly spaced, but still a convex grid, as in your code 
> example, you can use Gridded3DSet.


If you do need to use an Irregular3DSet, you can work around the bug 
with the following code:

    Irregular3DSet iset = new Irregular3DSet(xyz, samples,
      null, null, null, new DelaunayClarkson(samples));

Instead of:

>    Irregular3DSet iset = new Irregular3DSet(xyz, samples);


This technique works because for small samplings VisAD uses the 
DelaunayWatson algorithm by default, which fails with your data, but 
DelaunayClarkson works.

>    boolean success = iset.Delan.test(samples, true);


Also, please ignore the above Delan.test line; it was for debugging and 
I had meant to remove it.

-Curtis