I'm trying to understand how physics_model and its children are to be used to describe a collection of rigid bodies and constraints. I wonder if anyone can help by answering the following questions.

1. Does a rigid_body's <dynamic> child element specify whether the rigid body is movable with respect to the world, or only whether it is movable with respect to some containing physics_model?

2. Suppose a physics_model "A" contains rigid bodies B and C, but specifies no rigid constraints. Then are B and C (implicitly) linked rigidly, or are they completely unconstrained with respect to each other?

3. Suppose a physics_model "A" contains rigid bodies B, C and D, and specifies a rigid_constraint between B and C. Is component D then unconstrained, linked rigidly to B, or linked rigidly to C?

4. Suppose a physics_model "A" contains rigid body B and physics model C, and C in turn contains rigid bodies D and E. C also contains a rigid_constraint linking D and E. How can A specify a rigid_constraint linking B with, say, the D component of C? (Note that the description of rigid_constraint specifies that ref_attachment and attachment elements refer to rigid_body elements within the same physics_model, suggesting that a rigid_constraint in A cannot refer directly to either D or E.)

Thanks.