The technologies needed to build such a defense - rocketry, guidance systems, sensors and computers - have advanced spectacularly in recent years. And hitting a few missiles from a low-tech nation is far simpler than President Ronald Reagan's 1983 dream of a space-based defense that would make nuclear weapons "impotent and obsolete."
But the crucial question, often glossed over in the escalating international duels and domestic debates on the issue, remains the same as it was in 1944: Is it possible to shoot down an incoming missile?
Weapons scientists say the answer is still the same: probably not. "We don't have any great skepticism within government for the national missile defense, and we should," warned Jeremiah Sullivan, a physics professor at the University of Illinois. He is the chairman of a Department of Energy advisory committee on national security and the 2000 winner of the American Physical Society's prestigious Leo Szilard award.
Most of the 24 physicists, virtually all of them weapons specialists, interviewed in recent weeks said that even today's most sophisticated antimissile system could be easily tricked in its crucial last minute.
This is what they fear would happen to the system now being tested by the Defense Department's Ballistic Missile Defense Organization:
A minute after an enemy launched a missile attack on the United States, U.S. early-warning satellites would spot the missile's exhaust plume and activate the defense. Radars in California, Alaska and Massachusetts would begin tracking the missile and its cone-shaped warhead.
Roughly six minutes after the enemy launch, a U.S. rocket would blast off from Alaska. It would carry a 5-foot "kill vehicle" equipped with sensors, a telescope, an advanced computer, rocket thrusters and a guidance mechanism to steer the vehicle into a head-on crash with the 400-pound enemy warhead.
About 15 minutes into the warhead's flight, the two missiles would be closing on each other at a combined speed of more than 22,000 m.p.h., giving the kill vehicle less than a minute to find the warhead and adjust its course.
The world's most precise radar, yet to be built in Alaska, should tell the kill vehicle where to look. But in the final 30 seconds or so, it would be up to the vehicle's onboard sensors to lock on to the enemy warhead.
Today's best sensors can "see" many dots of light or heat 300 miles to 375 miles away.
But the incoming warhead could be spinning like a well-thrown football, tumbling end-over-end like a punt, or anything in between. In addition, the warhead could be accompanied by dozens of decoys.
If the kill vehicle's sensors choose the real warhead, the kill vehicle will blow the warhead into orbiting, without a nuclear detonation.
If the kill vehicle misses the warhead, one or more backup interceptors may be within range. But if all of the interceptors are tricked, it could be goodbye to a U.S. city.
"The current national missile defense really, in my view, is the worst of all possible worlds," said Wolfgang Panofsky, former director of the Stanford Linear Accelerator Center in Palo Alto, Calif., a Department of Energy physics research lab. "It's expensive. It's designed against only a small number of missiles. The countermeasures problem is fundamentally unsolvable. The system is simply incapable of dealing with very cheap decoys."
Seventeen of the 24 physicists interviewed, many of them recommended by the neutral American Institute of Physics based in Washington, do not think the Clinton administration's proposed National Missile Defense will work. Even scientists who support the proposal do not expect it to work every time.
In April, the governing council of the American Physical Society, the world's largest association of physicists, voted 34-1 to issue a policy statement saying the proposed missile-defense system is not technically feasible now because it cannot distinguish decoys from warheads. The lone dissenter thought the society should avoid political issues.
Still, the Clinton administration is poised to make crucial decisions about a technology that may not work. So far, the system has succeeded one out of two times in simplified - some say rigged - tests. During the first week of July, the U.S. Defense Department will conduct a key third flight test.
The real deadline is fall, when Clinton has promised Congress he will decide whether to move toward deploying a missile defense. The Ballistic Missile Defense Organization says it needs four years to have its Alaskan radar set up by 2005, when the North Koreans could have a nuclear warhead ready.
This rush-rush schedule worries Philip Coyle, the Pentagon's chief test officer, who warned in a report in February that Clinton would have to decide "based on a few flights with immature elements."
Rushed or not, the system is on track to fail, said Theodore Postol, a professor of science, technology and national-security policy at the Massachusetts Institute of Technology and one of the most outspoken critics of the missile-defense plan.
"It's got fundamental physical flaws that allow it to be defeated by simple countermeasures," said Postol, who is credited with proving that the Army's Patriot antiballistic missile failed most of the time during the 1991 Persian Gulf War. The Pentagon and the Patriot's manufacturer, Raytheon Corp., had claimed a 90 percent success rate.
In a May 11 letter to John Podesta, White House chief of staff, Postol said the new system would not work and charged that the early tests were "an elaborate hoax." Since that letter, which he also provided to the news media, was made public, two major contractors - Raytheon and Boeing Co. - have either declined to respond to news media questions or have referred reporters to the Pentagon's missile-defense organization.
In a written statement, the organization said Postol was "wrong" but declined to elaborate, "for obvious security reasons."
Shortly before Postol's letter was publicized, Lt. Col. Rick Lehner, the organization's spokesman, sounded cautiously optimistic about missile defense.
"At some point, it will be technologically feasible," he said. "We're not at that point yet."
A top missile-defense official at Raytheon, which is in line for $19 billion to build key radar and sensor systems and the U.S. kill vehicle, was more sanguine. Speaking on condition of anonymity and before the Pentagon stopped commenting, he said: "I don't need any technological breakthroughs to make this system work."
Nira Schwartz, a former TRW Corp. senior engineer, is dubious, however, because of tests she oversaw of a similar kill vehicle.
The July 1997 tests found that TRW's sensing system could not discriminate between decoys and the warhead, Schwartz said. Nevertheless, the Pentagon, TRW and their consultants all called the test a success. The TRW kill vehicle was not chosen but remains a fully funded backup to the Raytheon version.
Schwartz said in a telephone interview that she was fired in 1996 after telling TRW the system would not work. She was retained as a government consultant through 1999. She is suing TRW, alleging wrongful dismissal.
To circumvent the problem of distinguishing the warhead from decoys, the test results were manipulated to look successful, she said.
MIT's Postol wrote in his May 11 letter to Podesta: "In truth, the procedures followed by BMDO [Ballistic Missile Defense Organization] were like rolling a pair of dice and throwing away all the outcomes that did not give snake eyes and then fraudulently making a claim that they have scientific evidence to show that they could reliably predict when a roll of the dice will be a snake eyes."
The Ballistic Missile Defense Organization refused to respond to specific questions about the test, except to say that the allegations by Postol and Schwartz were wrong and that outside consultants had confirmed the test's success.
"The purported 'factual' assertions made by Ms. Schwartz and endorsed by Professor Postol are simply inaccurate and, apparently, are based on incomplete and misinterpreted information," TRW said in a written statement.
Another worry is making all the pieces of the complicated program work together. Small flaws can and do doom high-tech systems. Workhorse U.S. rockets, for example, have a 5 percent to 10 percent failure rate, and new systems have much higher rates of failure.
Edward Teller, 92, the father of the hydrogen bomb and a leading proponent of missile defense in the Reagan era, is not giving up.
"All we can ask for is a chance to be safe, and we should look for the best and most feasible chance," Teller said in a telephone interview. "The technology is a very difficult problem. But it is a problem that will not go away by saying no."
Seth Borenstein's e-mail address is sborenstein