optimization solvers in python

scikits has both 'optimization' and 'openopt' packages, though a cursory look at the repo for optimization (no docs) does not show anything that is obviously different from the simple stuff in scipy.optimize. openopt looks like a more extensive collection, put together by some ukranian academics.

cvxopt looks like the other big player in the python world. the convex optimization that it advertises is (i think) more geared toward complex constraints that i'm not currently worried about. also, openopt wraps at least some cvxopt solvers.

i need to take a closer look to decide which solvers to try, but it is probably a good idea to start with ralg for domain-contrained problems. that's the only one that doesn't require the initial point to be inside the domain.