Hi,

I am attempting a scenario where I could add 3D GeoNodes that are drawn as usual and viewed by a perspective cam, and GeoNode derivatives that are drawn in 2D screen coordinates viewed by a parallel camera. In the past, I hacked in to
void SceniXQGLSceneRendererWidget::paintGL() in SceniXQGLSceneRendererWidget.cpp and added
GLMatrixMode(GL_PROJECTION);
GLLoadIdentity();
GLMatrixMode(GL_MODELVIEW);
GLLoadIdentity();
to change change the coordinate and camera. And then I rendered the 2D on-screen geometry using good-old immediate mode glDraw issued within SceniXQGLSceneRendererWidget::paintGL().

Now I am much more interested in making the on-screen geonode more SceniX-ish, or, more scene-graph-ish. So I am thinking:

1) How do I extended the SceniX node family by sub-classing GeoNode? What specifics (e.g., object code) do I need to pay attention to? In 6.x there used to be a page within the chm help doc that describes this, but now I could not find any in the 7.x chm.

2) Where do I insert the GLLoadIdentity prior to rendering the on-screen stuff? Do I subclass the GL2 renderer and add a new handle method specialized for the oc of my GeoNode-sub-class?

3) How do I preclude the on-screen geometry from being included in the scene sphere / scene bounding box calculation, and prevent them from being traversed by the, say, stat traverser? Should I do it via the traverse mask?

I could not find pointers to the above questions in the help chm. Any suggestion would be gratefully appreciated!

Thanks!