Looking at the blog on collision filtering and considering how that relates to the lecture on ray-tracing, it seems trace and object channels will affect ray-tracing just as they do collision filtering. An example in the blog post noted visibility and penetrability as potential trace channel attributes, discussing how objects could be visible but impenetrable, visible and impenetrable, and invisible and impenetrable.
It also mentioned object channels which attach to an object and define how it responds to collisions when colliding with an object possessing a different object channel. As we consider making the grabber function work, object channels will allow relatively easy definition of objects that can be picked up and those that can’t. They could also be used to define which objects, while they can be grasped, either are or are not suitable for our purpose (e. g., is an object heavy enough to activate a trigger point).
Though I’m not yet sure how object channels are created, I can imagine they would be similar to components in that, once created, they can be applied to objects as appropriate. For example, if an object channel is defined as allowing an object to be grabbed and as being of a type that activates a trigger point, that channel could be applied to a chair, a table, or any other object the developer desire. Another channel could allow an object to be grabbed, but it might not be suitable to trigger.