I am having trouble finding a general technique that will allow me to efficiently and correctly hide objects in a space when a HUD (heads up display) is about to collide with the objects. My first naïve approach wrapped every object in the scene with a switch and a switching behavior triggered by WakeupOnViewPlatform[Entry|Exit]. This almost worked but had synchronization problems since the rendering and view platform movement happened on other threads. My second approach used a TransformGroup proxy placed between the view interpolator and the view platform. This allowed me to synchronize the view motion with object hiding. However, it meant that I had to manually detect and handle all corresponding "view platform entry and exit" events. I created a version that used picking that did seem to work, but it did not perform well with large numbers of objects and just seemed "too complex". My third approach uses a variant of the second approach where I handle all the boundary detection. However, this also does not perform very well and seems "too complex" for such a common function.
What other solutions am I missing? The first solution seems to place a reasonable burden on the application programmer, but the second and third solutions require so much work that I begin to wonder if I should bypass Java3D entirely since I have to manage and update the entire boundary system in the motion thread. Thanks for any information you can pass along. I am just a little frustrated! ==========================================================================To unsubscribe, send email to [EMAIL PROTECTED] and include in the body of the message "signoff JAVA3D-INTEREST". For general help, send email to [EMAIL PROTECTED] and include in the body of the message "help".
