Ismael is drinking tea in the dusty yard of the police station in Kajaki. The job has its advantages - such as meals of bread and rice twice a day - but the pay of $60 a month barely covers the price of his cigarettes. Nor has he been home for months - though his family lives only a few miles away. Kajaki is guarded by Royal Marines, but is surrounded by insurgents. "I will be disappointed if the British are beaten and leave," he says with some understatement. "The Taliban will kill me."
Spring will be crucial for him, as it will be for Afghanistan, for the 30-odd states engaged in rebuilding the country and, of course, for the region as a whole. Last year's fierce violence gave Nato countries a very nasty shock. The next few months will show if their hastily formulated plans to make up for the ground lost since the invasion of 2001 will work - or whether the violence will continue or worsen.
Few expect the 35,000 Nato troops - newly reinforced with American and British soldiers - to be defeated in the coming months; nor is President Hamid Karzai's government likely to collapse. But the fact that the Taliban are far from beaten is widely acknowledged. The outgoing head of the Nato forces, General David Richards, says the threat from the insurgents in Afghan istan has been "contained". Like Tony Blair, he insists that the war in Afghanistan is "win nable". Is it?
First, blame where blame is due. We are fighting a war that did not need to be fought. Travelling around the south and east of Afghanistan in 2002, I found much support for the western troops who, locals believed, had come to help them. Even in late 2003, in the small village south of Kandahar, where the Taliban had been founded nine years earlier, people had yet to turn against the coalition. But the west left the south of Afghanistan to rot. In 2004, three years after being invaded by an alliance of the richest countries on earth, there were malnourished children in Kandahar hospital. Apart from being a moral disgrace, the failure to keep, not win, the famous "hearts and minds" that are now so talked about, was criminally negligent.
But, despite all the mistakes, I am still optimistic. First, much of the country away from the south remains stable, relatively secure and shows considerable evidence of, albeit patchy, economic progress. This is a testament to the resilience and initiative of the Afghans.
Second, while the constant reliance on airpower remains a major problem - partly because of civilian casualties and partly because if the insurgents get hold of sophisticated surface-to-air missiles the whole operation will fall apart - at least the nature of the enemy has finally been recognised by the British army and many of its allies. The UK's strategy of trying to separate the "irreconcilable" Taliban hardcore from the "reconcilable" bought, coerced or merely disgruntled foot-soldiers is the right one. Some of the more enlightened soldiers have also grasped that "Taliban" is not necessarily the best label for the varied range of actors that comprise the rebellion, narco-war, jihad, jacquerie, fronde and revolt that is the war.
Third, it is now understood that progress in Pakistan - where an effective "state with no name" based on the Deobandi religious school network and attached political, commercial and military elements has been allowed to flourish over the past 30 years - is critical. And finally, most now see that the only people who can sort out Afghanistan are the Afghans, and that building Kabul's ability to deliver security and basic services and a form of democracy is better than getting foreign contractors to build bridges with other nations' flags painted on them.
So, five years too late, the west is getting a handle on what needs to be done. But recent summits aimed at getting more support for Afghanistan have been miserly affairs. Keeping the support and the interest of western populations is vital if Afghanistan's development is to be put back on track and then sustained.
For this, our leaders need to be realistic about what "winning" means. They need to jettison half the rubbish spouted by the government in 2001. That conflict was never about liberating women, developing Afghanistan or solving refugee crises - all of which could have been tried at any time over the previous decade - but about eradicating a security threat. We are not going to "liberate" women in Afghanistan for a generation or so, because trying to do so will play into the insurgents' hands. Every rash attempt to "reform" the ultra-conservative Afghan rural population in the past 100 years has sparked violence. Nor are we going to eradicate heroin. We are not going to stabilise the south rapidly either.
Instead, we can contain the insurgency while the rest of the country makes slow progress. But we need to be prepared to commit very considerable resources for a long time. And we need to have sensible expectations. In 20 years, if we are really lucky, Afghanistan will still be a fairly unstable, desperately poor, horribly unequal, often repressive and violent place - just less so than now. However, it will also have the chance to be none of those things a few decades further down the line. And that is the least we owe Ismael.
Jason Burke is European correspondent for the Observer and author of "Al-Qaeda: the true story of radical Islam" (Penguin)