North Korea is widely recognized as being years away from perfecting the technology to back up its bold threats of a pre-emptive strike on the United States. But some nuclear experts say it might have the know-how to fire a nuclear-tipped missile at South Korea and Japan, which host U.S. military bases.

No one can tell with any certainty how much technological progress North Korea has made, aside from perhaps a few people close to its secretive leadership. And it is highly unlikely that Pyongyang would launch such an attack, because the retaliation would be devastating.

The North's third nuclear test on Feb. 12, which prompted the toughest UN Security Council sanctions yet against Pyongyang, is presumed to have advanced its ability to miniaturize a nuclear device. And experts say it's easier to design a nuclear warhead that works on a shorter-range missile than one for an intercontinental missile that could target the U.S.

The assessment of David Albright at the Institute for Science and International Security think-tank is that North Korea has the capability to mount a warhead on its Rodong missile, which has a range of 1,280 kilometres and could hit South Korea and most of Japan. But he cautioned in his analysis, published after the latest nuclear test, that it is an uncertain estimate, and the warhead's reliability remains unclear.

Albright contends that the experience of Pakistan could serve as precedent. Pakistan bought the Rodong from North Korea after its first flight test in 1993, then adapted and produced it for its own use. Pakistan, which conducted its first nuclear test in 1998, is said to have taken less than 10 years to miniaturize a warhead before that test, Albright said.

North Korea also obtained technology from the trafficking network of A.Q. Khan, a disgraced pioneer of Pakistan's nuclear program, acquiring centrifuges for enriching uranium. According to the Congressional Research Service, Khan may also have supplied a Chinese-origin nuclear weapon design he provided to Libya and Iran, which could have helped the North in developing a warhead for a ballistic missile.

But Siegfried Hecker at Stanford University's Center for International Security and Co-operation, who has visited North Korea seven times and been granted unusual access to its nuclear facilities, is skeptical the North has advanced that far in miniaturization of a nuclear device.

"Nobody outside of a small elite in North Korea knows — and even they don't know for sure," he said in an emailed response to questions from The Associated Press. "I agree that we cannot rule it out for one of their shorter-range missiles, but we simply don't know."

Capabilities hard to assess

"Thanks to A.Q. Khan, they almost certainly have designs for such a device that could fit on some of their short or medium-range missiles," said Hecker, who last visited the North in November 2010. "But it is a long way from having a design and having confidence that you can put a warhead on a missile and have it survive the thermal and mechanical stresses during launch and along its entire trajectory."

The differing opinions underscore a fundamental problem in assessing a country as isolated as North Korea, particularly its weapons programs: Solid proof is hard to come by.

'North Korea understands that a serious attack on South Korea or other U.S. interests is going to be met with overwhelming force.'—Daryl Kimball, executive director of the Arms Control Association

For example, the international community remains largely in the dark about the latest underground nuclear test. Although it caused a magnitude 5.1 tremor, no gases escaped, and experts say there was no way to evaluate whether a plutonium or uranium device was detonated. That information would help reveal whether North Korea has managed to produce highly enriched uranium, giving it a new source of fissile material, and help determine the type and sophistication of the North's warhead design.

The guessing game about the North's nuclear weapons program dates back decades. Albright says that in the early 1990s, the CIA estimated that North Korea had a "first-generation" design for a plutonium device that was likely to be deployed on the Rodong missile — although it's not clear what information that estimate was based on.

"Given that 20 years has passed since the deployment of the Rodong, an assessment that North Korea successfully developed a warhead able to be delivered by that missile is reasonable," Albright wrote.

'Near suicidal' for NK to launch

According to Nick Hansen, a retired intelligence expert who closely monitors developments in the North's weapons programs, the Rodong missile was first flight-tested in 1993. Pakistan claims to have re-engineered the missile and successfully tested it, although doubts apparently persist about its reliability.

Whether North Korea has also figured out how to wed the missile with a nuclear warhead has major ramifications not just for South Korea and Japan, but for the U.S. itself, which counts those nations as its principal allies in Asia and retains 80,000 troops in the two countries.

Daryl Kimball, executive director of the Arms Control Association, acknowledges the North might be able to put a warhead on a Rodong missile, but he sees it as unlikely. He says the North's nuclear threats are less worthy of attention than the prospects of a miscalculation leading to a conventional war.

"North Korea understands that a serious attack on South Korea or other U.S. interests is going to be met with overwhelming force," he said. "It would be near suicidal for the regime."