I'm not on the beha
summarize the current consensus on the overall philosophy. I can
think of 3 approaches. Which, if any, of these are being considered?
(1) The scene graph lists all visible polygons as in current VRML.
30 times a second the rending engine takes exclusive control of the
graph and renders it to the screen. In between renderings, the
parallel beha
adding/deleting/moving/changing objects in the graph.
(2) The scene graph can have nodes that say "I don't know who my
children are, but interpret the following code and it will
recalculate my children". This means the rendering engine must query
every beha
objects aren't changing.
(3) The beha
with a semaphore guarding the screen while a task is using it.
In all 3 cases, I assume the beha
would have soli way to communicate with each other. I'm not asking
how that would work, I'm just asking how they communicate with the
screen. I would have assumed that (1) would be used because (2) would
slow down rendering, and (3) seems to have problems with one task
drawing a polygon on the screen, then another task overwriting it (or
one task erasing a polygon, and another task not knowing to redraw
the objects behind it).
Can anyone please summarize the current thinking is on this issue?
Leemon Baird
[email protected]
http://kirk.usafa.af.mil/~baird