CompuServe Thread

#Generic 3D -> 3DStudio

2 messages in this thread
#3885From: [F] Grant Blaha [Adesk]Oct 4, 1991 2:33 AM
Algorythmically, you are asking for a very difficult thing. A program would have quite a bit of trouble finding closed shapes in 3D space from an arbitrary number of lines with an arbitrary number of coincident points for each line. Consider a cube constructed of 12 lines, for example. First the program would have all the lines that are on a coincident plan, a rather time consuming task. Then the program would try to link up lines into closed regions, another task of trial and error. The next step the program would have to do is some 2D ray tracing (just like pattern hatching) on shapes that share the same plane to find shapes that are actually interior holes and do some checking to make sure none of the shapes intersect each other. Finally, the program would have to face the shapes. So, a program that would do what you suggest can be written, but it would be quite a project. It would be much easier for the translater in question to tesselate the polygons itself (AutoConvert does this for Generic 3D) or to at least write out the original polygon data as 2D polylines that contain all the above information except for the interior hole checking and face tesselation. Also, it would~ be easier for a program that already knows where the holes are to cut the shapes up in the translation, than for a program to look at the data after the fact. -GB-
#3905From: Steve OrmondeOct 4, 1991 10:19 AM
It does sound like more trouble than its worth. I'll stop beating you and Gary up on this and continue to beat up the cadkey folks. They move a lot slower than you folks (hence, the support is not near as good as yours). Thanks for you feedback. Steve