Questions on Baraff paper

Due February 19

1. Baraff says "Lagrange multipliers also allow us to handle nonholonomic constraints, such as
velocity-dependent constraints; reduced-coordinate approaches inherently lack this capability." Why do reduced-coordinate formulations lack the capability to handle non-holonomic constraints?

2. The paper gives a definition of a Lagrange multiplier. What is it?

3. Consider figure 1. Is the approach outlined in section five suitable for the structure in figure 1a? How about for figure 1b?

4. Is Baraff's approach suitable for figures 1a and 1b? Why?