Connections normally requires the attribute "ref" "positions" though maybe DX fills in the blanks if you don't provide it explicitly.
Rubbersheet would assume a 2D positions and scalar data to be used as Z. You don't want that as you already have Z in the positions. So you can just put your scalar data into "data", then Color/Colormap, etc. (and be aware of the "constantarray" object type: more compact in this case below, though not useful for your upcoming 'interesting' scalar data).
<x-tad-bigger>____________________________
Chris Pelkie
Vice President (607) 257-8335
Conceptual Reality Presentations, Inc.
30 West Meadow Drive
Ithaca, NY 14850
</x-tad-bigger>
On Jul 4, 2005, at 0:54, Bill Haneberg wrote:
I've looked through the archives and haven't found anything quite like the problem I've encountered, so I'm hoping that someone on the list might have a suggestion.
I am trying to plot a triangulated irregular network (TIN) representing the face of a rock outcrop. My approach has been to write a dx file defining a large number of triangles (all but the first values of each object omitted for brevity):
<x-tad-bigger>object 1 class array type float rank 1 shape 3 items 29406 data follows
1023.64400000000 5977.67500000000 22.3090000000000
< 29405 values omitted>
object 2 class array type int rank 1 shape 3 items 58785 data follows
1 12169 8620
< 58754 values omitted>
attribute "element type" string "triangles"
object 3 class array type float rank 0 items 29406 data follows
1.0
< 29405 values omitted>
attribute "dep" string "connections"
object "irregular positions irregular connections" class field
component "positions" value 1
component "connections" value 2
component "data" value 3
</x-tad-bigger>In the simplest case, I import the dx file and display the TIN using Import -> Color -> Shade -> Image. The problem is that the sloping irregular outcrop surface is bisected by a striped plane roughly parallel to the two long dimensions and perpendicular to the short dimension. There is also a large triangular artifact near the top of the outcrop. I know the triangulation is ok because the TIN plots without any problems in R and Mathematica. So, the problem is between me and OpenDX, not the TIN. I've put images of both the correctly rendered TIN and the problematic OpenDX version at this URL:
http://www.haneberg.com/tin.html
for anyone who would like to take a look.
I realize that I could Rubbersheet a collection of 2D triangles if my only objective were to display the surface. I eventually want to change the data values from the constant 1.0 to values that convey a scalar value for each triangle. I don't think that will work with Rubbersheet. Or, can I have one data component linked to the positions and a second data component linked to the connections?
Please let me know if you have any suggestions.
Thanks,
Bill
--------------------------------------
William C. Haneberg, Ph.D.
Consulting Geologist
www.haneberg.com
(206) 935-0846