A Mars colony doesn't have to be closed loop either. If you're bringing a large power source, you can make oxygen and fix carbon from the martian atmosphere. This is already part of the plan for the purposes of making fuel and oxidizer for the return trip.

Think of it more like a nuclear submarine with a greenhouse full of potatoes attached -- we can most likely do that.

Nobody's going to try to launch an entire mangrove forest on the first go. Biosphere2 is a failure of experimental design, not a failed experiment.

>Think of it more like a nuclear submarine with a greenhouse full of potatoes attached -

Maybe, but if we want to make Mars economically self sustaining the correct analogy is nuclear submarine that never surfaces and can totally self repair, including building new nuclear reactors and eventually replace all parts in itself.