Ok, running out the door, so I'll just touch one point:

The single essential task for stability in Afghanistan is not the "defeat" of any particular threat, nor is it the preservation of the current government. The later drives the former.

The single essential task is to change the government to a form that leverages the patronage culture toward stability rather than as a wedge that is continually driven between the victor and the loser. We must either:

1. Dedicate our main effort to forcing the current regime to take this on (and yes, the constitution is critical as it enforces this cultural wedge with the rule of law and Western support. If the design of the constitution does not describe and allow enforcement of a equitable distribution of patronage across society there will always be conflict. If the design of the constitution fuels a centralized upward Ponzi scheme of patronage there will always be outrageous, destructive corruption. If the constitution vests all power in one man, there will always be government in a box and never be legitimacy of government at any level in the eyes of people who are governed at those levels).

or

2. We must pack up and leave, recognizing that the only "vital interest" that we have in the region did not become vital until our own actions there had contributed so powerfully to the growing instability in Pakistan.

As an idealist, I vote for option 1. As a realist I vote for option 2. So, believe in 1, but execute 2. If we had vital national interests and if this were geo-strategically key terrain I would say commit to option 1 and make it work. We don't have either, so again the answer is 2.

What we do now is a threat-focused, tactically driven compromise that can only create suppression of threats for some temporary period of time to the current, unsustainable model of governance that is the root cause of the conflict. It is "doctrinally correct" but criminally stupid in its design.