I come from OpenSceneGraph which is C++ and have a different design approach. I'm thus a bit confuse as to how does one write custom Scenegraph node. In OSG each node is informed just before it is rendered and during the culling phase. In Xith3d it seems that all the rendering is done without giving control or any kind of feedback to the node.
In particular, I need to be able to get a hold of the viewing transformation (both model and projection) just before a particular node is to be rendered. I thought of creating a custom group node with one child (the actual Shape). But this fits the OSG model not the Xith3d.
So what is the correct way of acheiving this ? Do I need to write a custom shader ?