sherm1
commented
5 months ago IIRC this was an early project by @edrumwri that was superseded by later contact approaches in Drake. I think the LCP solver Evan wrote may still be useful but I'm not sure. Any thoughts @edrumwri and @amcastro-tri ?
Certainly we should remove the prominent discussion in the doxygen though since it describes an aspirational feature we abandoned.
As best I can tell, nothing in Drake nobody at MIT nor TRI is using the drake/multibody/constaint package, and yet we have docs featured prominently on our website like Multibody Dynamics Constraints that explain how we solve contact problems, and that seems to be to be grossly misleading since MbP doesn't actually do any of that.
Maybe I'm missing something, but I think we we should deprecate this module for removal? If none of us are using it, there's a fair chance there's a lot wrong with it or it will be confusing for people.
If anyone is still using it, they can just copy it to their own project. It's a standalone piece of code.
FYI It's not bound in pydrake.
@sherm1 Could you weigh in? Am I misunderstanding anything here?