I have a question about the position-based dynamics approach. Since the constraints are solved iteratively for every element in the simulation, you can still end up with a system where the constraints do not hold for some of the elements.
Example: you have a link between A and B, and a link between B and C. If you adjust A and B according to the length of the link, and then B and C according to the other length, you've moved B again, and the firsts link might not have the right length anymore.
Example: you have a link between A and B, and a link between B and C. If you adjust A and B according to the length of the link, and then B and C according to the other length, you've moved B again, and the firsts link might not have the right length anymore.