It would be a dream defense, which could destroy, in space, incoming nuclear or germ warheads.
Two years ago, the Pentagon agreed to let us watch an elaborate, expensive test of the missile defense system it is developing, a test using real rockets and sophisticated computerized technology.
It was estimated then that actually building the system might cost another $60 billion, on top of the tens of billions already spent.
Earlier this year, the Congressional Budget Office concluded the real cost may be much higher - over $230 billion.
To Pentagon critics, that's an enormous waste of money because they're convinced the system, as designed, is fatally flawed.
To the Pentagon, this missile defense system would be America's dream defense: a shield that would withstand virtually any strike, with more countries developing nuclear and biological weapons. The intelligence community believes a rocket carrying a nuclear or germ warhead could be shot at the United States within five years by North Korea or Iran, and a few years later by Iraq.
Lt. Gen. Ronald Kadish is in charge of building a ballistic missile defense system to defend the country against such potential enemies. The test July 8 was intended to be similar to a real missile attack. "It simulates all the things that have to happen in a combat situation," Kadish said.
The Pentagon's sophisticated and highly classified secret weapon for a real combat situation is called a "kill vehicle." It is designed to find and destroy the enemy warhead high above the surface of the Earth.
In the event of a missile assault, Pentagon radar systems are supposed to track the enemy warhead. Then the United States would launch the defensive rocket. In space, it would eject the kill vehicle, which would close in on the enemy warhead at a combined speed of 15,000 mph. It is called hitting a bullet with a bullet.
Kadish knew there was a lot riding on the July demonstration. "The test is about $100 million so we want to make sure that it counts," he said.
In an earlier test, the Pentagon destroyed a warhead in space, but critics claim that one was oversimplified and inconclusive.
And a previous test was a failure. The kill vehicle missed its target, providing more ammunition for the Pentagon's critics that this is an expensive, unworkable boondoggle.
The most outpoken critic, Ted Postol, said Kadish's system is doomed to fail. "Spending resources on doing serious scientific work on problems that are related to the ballistic missile defense problem is a perfectly appropriate thing for the United States to be doing," said Postol before the July 8 test.
"But we're not doing that. We're building things that have no chance of working instead," noted the physicist and MIT professor who was formerly a top U.S. Navy scientist.
The White House has called Postol arrogant, and even his colleagues say he is blunt and in your face.
But Postol does have a track record. In 1991, during the Gulf War, the Pentagon was claiming that its Patriot missiles were 90 percent effective in shooting down Saddam Hussein's crude but deadly SCUD missiles. After the war, Postol was the one concluding that the Patriots were nearly a complete failure.
"We analyzed at MIT the Patriot performance," explained Postol. "And our analysis indicated that the Patriots probably did not destroy a single SCUD warhead. Probably, the performance was zero."
After Postol's analysis, the Pentagon sharply lowered its estimate on the Patriot's performance.
According to Postol, the Defense Department is misleading the public again about missile defense. He said the stakes are much higher this time.
"Because if this system doesn't work, millions of people would die. This is a system that's supposed to defend people from nuclear attack. And if it doesn't work, lots of people would die," Postol said.
Since the early 1980s, said Postol, the Pentagon has accomplished very little in its effort to destroy enemy warheads in space - an effort that intensified when President Reagan talked about the initiative nicknamed "Star Wars."
During the Reagan years, the U.S. Defense Department went on a spending spree, trying to build a shield in space to defend against a massive Soviet nuclear attack. There were gadgets called brilliant pebbles to smash enemy warheads and ground-based lasers. Billions of dollars were spent on research but no effective missile defense system was ever built.
Now the Pentagon wants to funnel billions more into the new "kill vehicle" program. But there are reasons the system may not work. Before launching a rocket, an enemy can pack deflated balloons into it; later they are inflated and deployed with the warhead. The balloons camouflage the warhead or hide it; they can even be designed to completely enclose the warhead, making it virtually disappear.
"And these decoys are designed to make it difficult, or impossible, for the defense to understand where the warhead is relative to the decoys," Postol said.
And Postol believes if another country's military forces can reach the point where they can manufacture intercontinental ballistic missiles and the nuclear warheads to put on their tips, then it's safe to assume they can manufactue the decoys.
Out in space, the decoys and warheads look much the same, like distant points of flickering light, Postol said. And the infrared sensors on the kill vehicle couldn't be depended upon to tell them apart, he added.
"Although I can't see any feature, they're just a point of light, they might look a little brighter or dimmer," said Postol. "But the balloons are going to fluctuate in a way that's very similar to the way the warhead fluctuates. So the warheads and decoys all look roughly alike."
The professor said the Pentagon can do the difficult job of shooting down a warhead in space - of hitting a bullet with a bullet - but not if the warhead is surrounded by decoys. "If it can't tell the difference between warheads and decoys with a very, very high confidence, the system will collapse catastrophically."
Kadish wanted to prove that Postol was wrong with the July 8 test.
After midnight on July 8, at the underground command center at the Pentagon, it was almost time to launch a simulated nuclear attack on the United States.
And Kadish's job was to shoot that target down. If this had been a real attack the response time would be short, he said. "The decision makers...would probably have five to eight minutes to decide to enable the system."
Twenty-one minutes after the launch of the enemy rocket, it was time for Kadish's team to launch the second rocket, the defensive rocket with the kill vehicle.
"The interceptor launched and got off pretty good. So it's off to intercept. There it's going. And so we want to see it at a point in space where that 'kill vehicle' can open its infrared sensors and find the target and intercept," said Kadish as he monitored its progress.
The infrared sensors had to tell the difference between the warhead and the decoys. In earlier tests, several balloon decoys were used. But in this test there was only one decoy. "It's more than zero," said Kadish. "And just as we don't go supersonic on our first flight test of an airplane; we want to take this a step at a time."
The Pentagon's critics say the sensors are so essential to this system that tests are useless unless Kadish can prove the sensors can discriminate between the decoy and the warhead.
But as things turned out, Kadish was not able to even test those much-criticized sensors. First, the one balloon decoy, designed to confuse the kill vehicle, did not inflate properly.
"So the decoy is not going to look exactly like what we expected. It presents a problem for the system that we didn't expect," said Kadish.
The general wasn't happy, but a few minutes later, he had an even bigger problem. The kill vehicle was still attached to its booster rocket - unable to separate for some reason - and therefore was unable to even try to intercept the enemy warhead.
If this had been a real attack, the warhead would have continued on to its target.
The July 8 test failed more fundamentally than even Postol could have imagined. But to him, the missile system is just another in a long list of failures dating back to tests in the Reagan years.
"In Star Wars, we were talking about X-ray lasers, and they didn't work," said Postol. "We were talking about deuterium fluoride space-based lasers. They didn't work. We were talking about hydrogen fluoride lasers in space, and they didn't work. We were talking about neutral particle beams, and they didn't work. We were talking about charged particle beams, and they didn't work, just went on and on and on. Now we're down to interceptors, and they don't work."
And Postol said the Defense Department has known that for years. One woman was warning the Pentagon back in 1996 that a major defense contractor was lying when it said the infrared sensor technology did work, Postol said.
In the last two years, the Pentagon has kept on testing.
They've conducted three more tests, all with real rockets and decoys, and the Pentagon says every one of them was successful.
But Postol says it's all smoke and mirrors. He agrees that the incoming missiles were destroyed in all three tests… but Postol also says the Pentagon deliberately designed the decoys so they could not confuse the kill vehicles - and so the tests could not possibly fail.
Find out what this early critic had to say in A Far-Off Dream?