From an article in today's Washington Post by former Clinton Defense honchos Aston Carter and William Perry:
Should the United States allow a country openly hostile to it and armed with nuclear weapons to perfect an intercontinental ballistic missile capable of delivering nuclear weapons to U.S. soil? We believe not. The Bush administration has unwisely ballyhooed the doctrine of "preemption," which all previous presidents have sustained as an option rather than a dogma. It has applied the doctrine to Iraq, where the intelligence pointed to a threat from weapons of mass destruction that was much smaller than the risk North Korea poses. (The actual threat from Saddam Hussein was, we now know, even smaller than believed at the time of the invasion.) But intervening before mortal threats to U.S. security can develop is surely a prudent policy.
Therefore, if North Korea persists in its launch preparations, the United States should immediately make clear its intention to strike and destroy the North Korean Taepodong missile before it can be launched. This could be accomplished, for example, by a cruise missile launched from a submarine carrying a high-explosive warhead. The blast would be similar to the one that killed terrorist leader Abu Musab al-Zarqawi in Iraq. But the effect on the Taepodong would be devastating. The multi-story, thin-skinned missile filled with high-energy fuel is itself explosive -- the U.S. airstrike would puncture the missile and probably cause it to explode. The carefully engineered test bed for North Korea's nascent nuclear missile force would be destroyed, and its attempt to retrogress to Cold War threats thwarted. There would be no damage to North Korea outside the immediate vicinity of the missile gantry.
The U.S. military has announced that it has placed some of the new missile defense interceptors deployed in Alaska and California on alert. In theory, the antiballistic missile system might succeed in smashing into the Taepodong payload as it hurtled through space after the missile booster burned out. But waiting until North Korea's ICBM is launched to interdict it is risky. First, by the time the payload was intercepted, North Korean engineers would already have obtained much of the precious flight test data they are seeking, which they could use to make a whole arsenal of missiles, hiding and protecting them from more U.S. strikes in the maze of tunnels they have dug throughout their mountainous country. Second, the U.S. defensive interceptor could reach the target only if it was flying on a test trajectory that took it into the range of the U.S. defense. Third, the U.S. system is unproven against North Korean missiles and has had an uneven record in its flight tests. A failed attempt at interception could undermine whatever deterrent value our missile defense may have.