The question is still Palkistan. Al-Qaeda's plans have always extended past Afghanistan to Pakistan. The warlike response to 9/11 and US diplomacy involving India as a pressure lever set those plans back. But al-Qaeda's plan for Pakistan has continued to execute, and is now close to success.

While the realist strategy of relying on a friendly dictator has again been proven fatally unstable, it's worthwhile to note that many of al-Qaeda's gains in Pakistan came during Bhutto's tenure as well. The utterly corrupt "Mr 10%" and his PPP party show very little enthusiasm for taking on Pakistan's Taliban, which means they're going to lose to it in the medium term. Al-Qaeda is marshaling the resources and population of western Pakistan, which they rule absolutely via a series of Emirates/fifedoms that the army is very reluctant to enter, and finance it all with poppy trade profits.

Ah, the poppy trade. Which, as Lt. Col. Krenson reminds us, is something that can't be stopped unless NATO wants to hurry things up and lose Afghanistan immediately.

The ultimate question is whether even an Afghan government whose Army is fully trained and deemed to stand on its own can successfully hold out against al-Qaedistan and its resources in Pakistan, operating from a safe haven. I doubt very much that it can. And NATO isn't about to stick around in perpetuity to fight a war that is intrisically unwinnable because Pakistan is a perpetual safe haven under partial or full al-Qaeda command.

Are things better now than they were? Yes. Does something big have to change, in order to avoid failure in Afghanistan? Yes. Is the situation unstable enough that it's worth sticking it out another couple of years to see if events in Pakistan force a decision or create a big civil war that lets us fight al-Qaeda more directly there (or forces far more drastic measures)? Probably.

But... and this is unpalatable, but true. Iraq is strategically winnable - and Afghanistan, unless something big changes, is not.