This past week has seen appalling terrorist violence in Pakistan and Afghanistan. The bombings in Pakistan were designed in part to coincide with the visit of US Secretary of State Hillary Clinton. As always, these bombings were designed to kill, but they were also designed for the evening news in every Western country that has troops in Afghanistan and a stake in Pakistan.
The war is going very, very badly in both countries. Meanwhile, the whole world waits for yet another US review of its Afghanistan and Pakistan policy.
To simplify rather drastically, the two possible alternatives are counter-insurgency, which often travels under the acronym COIN, and counter-terrorism, or CT.
COIN is advocated by General Stanley McChrystal, US commander in Afghanistan. He wants 40,000 more US troops. Right now there are 68,000 Americans in Afghanistan, 35,000 from other nations and roughly 170,000 Afghans divided between the police and army.
CT is advocated by US Vice-President Joseph Biden. It is based on the idea of lowering the number of US troops in Afghanistan and concentrating the US effort on killing terrorists as they emerge.
President Barack Obama has a mandate to win in Afghanistan. All through the election campaign he promised to give the Afghanistan campaign the resources it needed.
He then held a policy review after he came into office and declared in March: "To focus on the greatest threat to our people, America must no longer deny resources to Afghanistan because of the war in Iraq."
But as Der Speigel asked, can a Nobel Peace Prize winner wage and win a war in Afghanistan? The Obama administration seems overwhelmed. It is simultaneously dealing with healthcare reform, the fallout from the financial crisis, the Afghanistan/Pakistan disaster and the demand for a global warming agreement.
One response of hard-pressed leaders is to commission further study, so that while that's under way they can concentrate for a time on one of the other pressing issues. But while understandable, that's nowhere near good enough for Afghanistan and Pakistan.
At the moment the allies look, perhaps not entirely accurately, as though they are losing in Afghanistan, while Pakistan increasingly appears caught in a monstrous civil war that will challenge every institution in that fragile, broken-backed society.
To try to understand what's going on, it's helpful to disaggregate the forces at work. In Afghanistan, the US-led coalition, of which Australia is part, in alliance with the government of Hamid Karzai, is fighting the Taliban.
This Taliban is made up of several different forces. There is the central, and profoundly ideological, group led by Mullah Omar, the former head of the Taliban government. There are regional Taliban movements as well, some of which were in government with Omar. Beyond this, smaller tribal groups and clans have made alliances of convenience with the Taliban. Some elements of the Taliban are less ideologically committed than the Mullah Omar group.
The top Taliban leadership base themselves in the tribal areas of Pakistan.
At the same time, there are now Pakistani Taliban, whose leadership is also based in Pakistan's tribal areas. They, too, are increasingly allying themselves with other Islamist movements within Pakistan. The Afghan Taliban were initially sponsored by the Pakistani military, who have also sponsored other Islamist extremist groups, mainly to attack India.
However, the Pakistani state is in danger of being eaten by the monsters it created. The recent wave of attacks against Pakistani military bases shows there is now an all-out war against the Pakistani state by the Pakistani Taliban.
The Pakistani military, having recently retaken control of the Swat Valley, is now involved in a massive, anti-Taliban campaign in Waziristan.
One of the world's foremost experts on Afghanistan and Pakistan, Anthony Cordesmann, told me this week he's sceptical about what the Pakistani military will be able to achieve in Waziristan. "Tactically, they'll be able to smash their way in all right," he said. "But it's another question whether they can clear, hold and build. They are basically a flat-land army designed to deal with India. They have some heavy learning experiences ahead of them.
"Whether they can adapt and learn effectively is the question."
There is some consolation to be had from the fact that the Pakistani military now sees the Pakistani Taliban as unambiguously its enemy, and the enemy of the Pakistani state.
Pakistani soldiers are infinitely more likely to be effective fighting for their own country, than they are in meeting international obligations to police international terrorism, where perhaps they don't see their own interests fundamentally at stake.
However, the Pakistani military has still not severed its links with the Afghan Taliban, which it believes might come back into power in Afghanistan and which might, in Pakistani eyes, rule Afghanistan in a way which is compatible with Pakistani interests.
This is so even though the Afghan Taliban and the Pakistani Taliban are in broad alliance and frequent contact.
So where is al-Qa'ida in all this? The best intelligence guess is that al-Qa'ida's leadership is also headquartered in the tribal areas of Pakistan. Al-Qa'ida has a symbiotic relationship with the Afghan Taliban. Those who favour the Biden CT approach against the McChrystal COIN approach often argue that it should be possible to detach al-Qa'ida from the Taliban, and fight the former and make a deal with the latter.
The problem is there is very little historical evidence that this can be accomplished. After the 9/11 terror attacks, Washington gave the Taliban government in Afghanistan every chance to give up Osama bin Laden or at least expel him. Even though Mullah Omar knew that sticking with bin Laden could see his government destroyed and his rule over Afghanistan ended, he did stick.
Since then, if anything, the relationship between al-Qa'ida and the Taliban has grown closer. Al-Qa'ida has trained the Taliban in every terror trick they know, so that Taliban insurgent operations have become ever more sophisticated. They also acknowledge the Taliban's leadership. At the same time the Taliban continues to provide hospitality and support to al-Qa'ida. Osama bin Laden may move around a lot, but he almost certainly isn't hiding in caves. He is staying as an honoured guest with old and deep friends.
Stephen Biddle, who was a member of McChrystal's assessment team, has written a devastating critique of the CT approach as a way of lessening the US troop commitment. He summarises McChrystal's COIN approach as being focused on protecting the Afghan population, expanding its army and police, reforming government, providing economic development, weaning Taliban fighters away from Mullah Omar and targeting those who refuse. To do this effectively requires doing it all, and it requires more resources.
Biddle goes through the alternative approaches of CT. One is: train the Afghans, don't fight on their behalf. This won't work, he says, because effective training effectively requires more US troops. The only really effective training involves mentoring by integrating coalition troops with Afghan troops in battle. This requires a lot of coalition troops.
Another suggestion is the greater use of unmanned aircraft to attack al-Qa'ida leaders. But to be effective this requires human intelligence which is only available from a sympathetic government and a large presence on the ground. Yet another is to buy off warlords. This is indeed also part of the COIN strategy, but the warlords won't stay bought if they think the US and its friends are losing or withdrawing. They'll take coalition money and then join the enemy when it turns up in force anyway.
Another line favours sending civilian aid rather than troops, but no aid project survives in contemporary Afghanistan without security protection. The Taliban will never allow civilian aid to prosper if it has the power to obliterate it. The final piece of CT advice is to tread softly, because having too many foreign troops annoys the Afghans and creates a bigger backlash. But tread softly was Donald Rumsfeld's policy and it got Afghanistan into the mess it's in today. There are enough foreign troops already to annoy a lot of Afghans, but not enough to provide security.
The situation has been vastly complicated by the corrupt presidential election and the loss of credibility for Karzai's government. And the polls are bad for the Afghan war in America. But this is exactly when presidential leadership is most needed. A Taliban victory in Afghanistan would put huge pressure on nuclear-armed Pakistan, empower al-Qa'ida terrorists and could well see Taliban-style terror armies replace al-Qa'ida as the jihadist modality of preference, such that similar groups emerge in Central Asia and even other parts of South Asia.
It is overwhelmingly in US, and Australian, interests for this not to happen. Whatever strategy the US adopts must be coherent and resourced to succeed.
The world continues to wait, and wait, for Obama to make up his mind.