StrategyPage discusses US plans for Afghanistan:

It would take about 18 months to get all the new forces to Afghanistan. That would then result in a Western force of about 100,000 troops (62,000 U.S. and 40,000 NATO). In that time, the Afghans are expected to expand their own security forces (police and army), arm and train some tribal militias, to produce a total force of nearly 300,000 local and foreign troops and police.

Canada, and some other NATO members, object to the U.S. plan to provide weapons and training to form more reliable anti-Taliban militias. The U.S. and Britain believe these militias are an acceptable risk.... Many NATO nations are appalled at the amount of corruption in Afghanistan, with tribal leaders often keeping most of the aid provided to their tribe, for themselves. These nations prefer to put more effort into cleaning up the government, police (which are notoriously inept and corrupt) and improving the army (which is pretty good, but small). But the Americans and British have worked with these tricky tribal situations often in the past. As the Brits like to put it, "who dares, wins." The Americans have decades of experience with the tribes, having been there since the 1980s, during the Russian invasion of Afghanistan. Europeans don't always trust American combat experience, it being an article of faith in Europe that the U.S. invasion of Iraq was a mistake, and winning that war was a fluke. Those lucky, but ignorant, Americans! The anti-militia crowd will lose this argument, but you'll see a lot about it in the media. Allies squabbling, whether real or imagined, makes for exciting news. But the basic American strategy is to play the game the Afghan way. The tribes are in it for the long haul, and will change sides if they sense they are losing. The Americans and Brits want to use superior firepower, mobility and cash reserves to flip as many tribes as possible, as quickly as possible, to hand the Taliban a very obvious, and well publicized "defeat."

I still think Pakistan is the key problem, owing to sheer numbers. Still, this strikes me as the best approach within Afghanistan. Hopefully, the interim strategy will also take some pointers from this source.