Re: Gridded3DSet

• To: John Yao <johnyao@xxxxxxxx>
• Subject: Re: Gridded3DSet
• From: "From: Bill Hibbard <hibbard@xxxxxxxxxxxxxxxxx>" <billh@xxxxxxxxxxxxx>
• Date: Thu, 3 Jan 2002 15:44:06 -0600 (CST)

Hi John,

> I have created an ellipsoid using the Gridded3DSet.  However, in order to 
> create proper grids, two functionTypes and two Gridded3DSet were need; one 
> for the top hemisphere and one for the bottom.  The incomplete code look 
> something like the following:
> 
> functiona = new FunctionType(xyz, v); //function domain as xyz and range as 
> v (for surface mapping).
> functionb = new FunctionType(xyz, v);

Since these FunctionTypes are identical, you only need one
of them.

> Gridded3DSet seta = new Gridded3DSet(xyz,samplesa, 41, 41);
> Gridded3DSet setb = new Gridded3DSet(xyz,samplesb, 41, 41);
> 
> This works fine and the ellipsoid looks good, but there is a very obvious 
> cut in the middle (like an equator on a globe).  Should I try another way 
> of rendering this?  Is there a way to do this without two sets and two 
> functiontypes?

Yeah, you can make a 2-D grid (in a Gridded3DSet with manifold
dimension = 2) where the four edges are all very close to the
north pole.

If you use an Irregular3DSet with manifold dimension = 2, the
samples can be more evenly spaced, although it is a little tricky.
You have to create an appropriate DelaunayCustom for the 2-D
topology.

> On a separate issue:
> I was thinking of using irregular3DSet, but I don't fully understand it.
> Just to test irregular3DSet I replaced the Gridded3DSet as:
>          Irregular3DSet seta = new Irregular3DSet(xyz, samplesa);
>          Irregular3DSet setb = new Irregular3DSet(xyz, samplesb);
> But all I get was points in space.

You need to use the constructor where you pass it a DelaunayCustom
(extends Delaunay) with manifold dimension = 2.

Good luck, Bill

• References:
  • Gridded3DSet
    • From: John Yao