Re: Livegraphics3D

*To*: mathgroup at smc.vnet.net*Subject*: [mg18389] Re: Livegraphics3D*From*: Martin Kraus <Martin.Kraus at informatik.uni-stuttgart.de>*Date*: Wed, 7 Jul 1999 00:10:56 -0400*Organization*: University of Stuttgart*References*: <7kpb8d$4am@smc.vnet.net> <7lcebc$1iv@smc.vnet.net>*Sender*: owner-wri-mathgroup at wolfram.com

Dan Truong wrote: > > I am using LiveGraphics3D too, > > The problems you have I think come from the imprecision of the > LiveGraphics3D drawing engine. Looks fine otherwise. > > When surfaces intersect, result is novey satisfactory... Same when > the viewing angle overlaps 2 polygons which are close to each other > (I believe an error occurs when 1 of the front corners which should > be hidden for a polygon is at the same or lower depth than some of the > corners of the masking polygon). > > Any comments Mark? > > Dan- Who is Mark? ;) You are perfectly right, Dan. Actually these are the problems of the simple Painter's Algorithm used in LiveGraphics3D. Any alternative I am aware of (like z-buffering) would need too much performance in a Java implementation. One way to overcome these problems is to avoid the intersections of polygons with Mathematica by breaking all polygons along their intersections (as was done for many of the polyhedra in Erics Treasure Trove of Mathematics), however, any automatic procedure does not work well for most surfaces. Thus, finally, I suggest to either split up the surfaces into smaller pieces, minimizing the unwanted effect, or to break polygons along the line of intersection, which might involve quite difficult programming. Greetings MARTIN Kraus :)