[FreeWRL-develop] X3D over collada

[doug s. <hig...@ho...>]

GoogleEarth: KML + DAE
3DWarehouse-friendly web3d: X3D + DAE
GoogleEarth shows .KML files -which have geographic coordinates, and some declarative animations such as Photo/Camera animations for photoTours and spherical panoramas- and also KML can render a collada file by refering to it with an href
     <Link>
      <href>bfiles/Garage.dae</href>
     </Link>
as I did for this project www.airdrie.ca/cent/google 
Collada uses rectilinear coordinates. The KML file positions it geographically, and the .dae contains local coordinates in meters by default IIRC.
sketchup.google.com/3Dwarehouse stores collada files. Not KML. So a content developer would download a .dae model, and then position it in googleEarth. There's also a process in Sketchup for positioning geographically, and in that case you would SaveAs .kmz file which includes both kml and collada.
KML has a few delcarative animations - such as camera animations for a photo tour. But no scripting, no general animation.
Something similar could be done for X3D: X3D refering to collada as passive geometry. But how to animate the geometry? For example if I download a vehicle in collada format, how from X3D would I animate the doors to open on their hinges when clicked?
One idea is to use the web3d 'TransformSensor' node in combination with a viewer than could use it on the .dae child geometry. I could hook 4 TransformSensor nodes to animations for opening vehicle doors, and the transformSensor boxes spaced out like vehicle doors, and perhaps wrap it in a ProtoDeclare. Then in a special editor, place the .dae vehicle. Then drag an instance of my DoorOpener proto to overlap the vehicle. Then scale the DoorOpener proto to the vehicle. If I have a cooperative viewer, it would then detect the door geometry inside the TransformSensor box, and allow me to click to open the door.
The goal with this design: to keep animations in the upper .x3d level. So static geometry can be added and replaced as .dae without needing to mix the two. Keeping the animation layer separate allows for quick updates of geometry from good geometry authoring tools like sketchup, and from the 3Dwarehouse for collada. While keeping the web3d strength -scriptable animations- in a separate layer, much like google did with KML and geographic coordinates.
-Doug