The Royal Society of London has released the results of a year-long study into various geoengineering options, and have looked at them in terms of cost, risk, and so forth. More on the report can be found here (also, the 2020Science blog has some good posts on it as well).
I must say, I agree with the need to further consider these options. Yes, we need to look at mitigating emissions, of course, but if Kyoto is any indication, we stand the real possibility of failing to reach the various global and country-specific targets to be set in the upcoming year or so. And it is highly likely that new technologies are yet to be invented which will help in these efforts. But a plan B is generally a good idea - even if we never implement such measures, better to plan for them now than, in 10+ years’ time, to realise that such measures are necessary, but cannot be implemented in a timely fashion.
Also, there’s the pure geek joy of simply contemplating huge mirrors on the sky (having said that, Futurama has already warned us of the potential pitfalls of such a technology).