On Wednesday, Oct 20, 2004, at 17:37 America/New_York, Matthias Kraatz wrote:
There must be a solution to this. I am sure there are more applications like this which make use of geometry data that typically come in faces, edges & loops format and data maps on grids.
Unfortunately, I don't agree with this assertion. In my 13 yrs of DX use, I've used FLE representations twice, and neither time for data mapping problems. Once as polygons (political boundaries), that were colored uniformly over each entire polygon, and currently, in a project that represents materials as polycrystals (arbitrary polyhedra). We can assign colors by grain ID or other simple scalar function to color all surfaces of each polyhedron. There are advantages to using FLE reps for speed and reduced data download overhead from the database where the data objects reside.
However, to put physical/mechanical data fields onto these crystals, the research group also has meshing algorithms to triangulate the surfaces and tetrahedrally subdivide the volumes, providing the far more DX-compatible interpolation elements of triangles and tets. This makes much larger objects but we are developing subselection routines to reduce the download of unneeded data, such as letting the user choose to focus on a subsection of interest before calling for the mesh to download.
So I would say that you have to attack the problem by triangulating the irregular mesh, not use FLEs. I don't know whether you can use the same mesh for the other data field, or if Map will give sufficiently good interpolation.
FLEs are IMO, the black sheep of the DX object family; occasionally useful, but not full-fledged members.
_______________________________
Chris Pelkie
Scientific Visualization Producer
622 Rhodes Hall, Cornell Theory Center
Ithaca, NY 14853
