The Rand Institute has a study out which essentially recommends a return to the Clintonian system of treating terrorism as a criminal problem, and not a military one. Without going into various theories about the bogeyman, Israel, and oil multinationals, this seems to make rational sense.
Here's what sorts of tilts my pinball machine a bit: check the pie chart. According to the database Rand used to draw some of its conclusions, 53% of the terror groups disbanded in the last fifty years did so on the basis of either political compromise, or outright victory. 53% of the time, the terrorists walk away with at least some of what they wanted.
If terrorism 'works', that says a lot about how the nations affected by it are failing to achieve a consensus of the governed. Given that what a potential terrorist wants is not unconscionable or impossible, (ie, if you're willing to give in anyway) it seems both morally correct, and highly cost effective to make the necessary concessions quietly, and in advance of the actual terror.
Then again, 53% isn't an overwhelming margin of victory. I'd be very curious to see a table of demands, cross referenced to likelihood of victory. What got won? Was it just money? Separatism? Something a western nation might sympathize with? The implications seem critical given the last eight years of failure.