Consider the notion of shading the planet with mirrors. The US National Academy of Sciences found that 55,000 orbiting mirrors would reflect enough sunlight to counter about half the doubling of carbon dioxide.
But each mirror must be 100 sq km; any larger and you would need a manufacturing plant on the Moon, says Dr MacCracken. The price tag of space-based fixes makes them prohibitive - for now.
The biggest problem with that approach, once you've figured out a way to build and put the shields in orbit, is to get the balance right ("current computer models are not up to the task of predicting the consequences of large-scale plans such as Earth shades."). It's hard to predict the side effects of reducing the amount of solar light that hits the Earth, though it is probably easier (relatively) to fix unexpected problems by adding or removing solar shields than it is to affect the atmosphere in other ways.
I am not saying that this is the kind of thing that we should focus our energies on, and I would be extremely sad if that kind of geoengineering hack was hijacked by those who want to maintain the status quo (for financial or ideological interests). As David Keith said: "The knowledge that we maybe could engineer our way out of climate problems inevitably lessens the political will to reduce emissions." But it's another tool in the toolbox, not to fix the problem, but maybe to buy some time if things go from bad to worse faster than expected; buy some time while we work on a more permanent solution.