Kabul, Afghanistan ? It is cold and wet on the vast and desolate grounds of the Kabul Military Training Center, ringed by mountains on the outskirts of the city. Here, Afghan officers backed by NATO mentors are training new recruits to shoot and care for their AK-47s.

Seventy percent of the new grunts are illiterate, although officer cadets have a high school education. But the motivation of these youths seems high, in a country where the Afghan National Army is a respected institution.

On the shooting range, trainee Mohammed Arif from Kandahar has no qualms about the possibility of fighting a fellow Pashtun who is a Taliban. “If he comes to destroy my country, then I will kill him because he is the enemy,” he says.

And how do these young men feel about the increasing U.S. military presence in Afghanistan? I ask Ubaid Allah, another Pashtun recruit from Kandahar, where most of the new U.S. troops are headed. His swift reply: “They are welcome since they are coming to train us. When we can stand on our own feet, they can leave, but if they leave now, the Taliban will be downtown.”

Here in a nutshell is the essence of the U.S. military’s clear, hold and build strategy in Afghanistan: Push back the Taliban (clear), while training a larger Afghan army to help hold the security gains. Meanwhile, help a weak Afghan government deliver development aid to its people (build), so unemployment and poor governance won’t drive the populace toward the Taliban.

The many difficulties of this strategy are very apparent. America’s postwar shift in focus from Afghanistan to Iraq permitted the Taliban to regroup and link up with Pakistani Taliban and al-Qaida in Pakistani safe havens. A weak Afghan government — headed by President Hamid Karzai, who is likely to be re-elected in August — has failed to deliver jobs and services to its people.

Yet, despite the hurdles, I was impressed by the realism of U.S. military commanders. “At the six-year mark here we realized we were screwing up and at the seven-year mark we finally started getting it right,” said Col. John Agoglia, director of the U.S. counterinsurgency training center in Kabul.

Getting it “right” means an understanding that military means alone won’t resolve the problem. “We’ll never stop the insurgency by killing Taliban,” Maj. Gen. Michael Tucker, deputy commander of U.S. forces in Afghanistan, told me. “Only the local population will stop it.

“We’ve got to give the people … a better alternative to the Taliban,” Tucker said. “The people are not mesmerized by the Taliban at all. They are thirsty for good governance.”

Indeed, polls show the popularity of the Taliban, which terrorizes border regions, assassinates government officials and burns girls’ schools, is measured in single digits. But a combination of intimidation and cash, much of it from the poppy crop, allows the Taliban to continue advancing. Talk of talking to Taliban leaders (as opposed to peeling away lower-level commanders) is pointless when those leaders think they are winning.

So the NATO strategy, says Tucker, is to focus simultaneously on increasing the size of Afghan security forces, while encouraging “a civilian surge of development. One alone won’t do it.”

That means an intense push to train and mentor Afghan military and police. Meantime, U.S. commanders and diplomats are exploring better ways to channel American aid and technical expertise directly to projects that are really needed by local communities.

One idea: In troubled Kandahar province, U.S. commander Brig. Gen. John Nicholson is talking with Afghan officials about funneling money to projects vetted by elected community development councils. The aim would be to avoid long bureaucratic delays and foreign contractors, and give Afghans a sense of ownership over the projects. The general also talks of helping Kandahar farmers cultivate substitute crops for poppy. “The idea,” he says, is to use these development efforts “as a catalyst to bring (Afghan) government together with the people. This is crucial.”

U.S. commanders are also sensitive to the need to reduce civilian casualties, while recognizing some collateral damage is unavoidable. “We’ve made a great effort to fix the DNA of inbound units before they get here,” says Gen. Tucker. “I say to them, ‘If you think your mission is to kill as many Taliban as you can, don’t even bother coming.’ Your job is to connect with this population here.”

In other words, the U.S. strategy is population-centered rather than enemy-centered. In principle, this is the only strategy that has a chance to check the Taliban. In practice, there is still a very long, hard way to go.

— Trudy Rubin is a columnist for the Philadelphia Inquirer.