The final time weighting isn't correctly implemented in the boundary conditions, it should be the derivative rather than the function itself.

References have be made to weightings in the optimal control formulation. Firstly, the final-time penalty has yet to be correctly implemented and secondly the weightings have actually been included. Weightings either need to be fully included or fully removed.

The whole project needs to be better documented, from comments to README.

Both the index-3 example and distillation model result in a singular collocation matrix. As far as I am aware both models have been correctly implemented, but this hasn't been independently verified. It could be be a feature of the collocation itself, although this is also yet to be theoretically justified.

The scikit branch uses the scikit.bvplg1 python wrapper for colnew rather than ScipPy's bvp solver, so the installing script needs to reflect this dependency.