> It's become clear that the direction we want to take with Annotation text of
> the sort Len wants is to make nodes which orient themselves fixed with
> respect to the camera. We have yet to design this node type; one issue in
> particular that was raised was how a camera-oriented node behaves when the
> viewpoint jumps to another camera.
At the moment, the Webspace browser already allows for camera-oriented
nodes. However, they cannot be actually selected as hyper-link at the
moment.
If you have more than one camera node, within separate separators,
only the first camera gets transformed when you manipulate the view.
(The two cameras render their objects on top of each other,
appropriately Z-buffered.) Maybe it's a side-effect, but the Webspace
browser already does what we want, without any additional nodes. On
second thought, it doesn't allow for screen-aligned annotation text,
just for screen-aligned floating interfaces, but it seems to me to be
enough of a "close-enough" functionality that we might be able to pull
it into the 1.0 spec.
Tom Meyer
WAXweb
Brown Graphics Lab