Back in the early 1980s, after some scandals involving major weapons systems that were deployed before they'd been adequately tested, the Pentagon adopted a policy of "fly before buy." No system would graduate from research and development to procurement until it had passed a series of tests—not just "development tests," in which a weapon demonstrates certain technical milestones, but "operational tests," in which the weapon demonstrates it can do what it's supposed to do in an environment simulating combat.
The interceptors at Fort Greely haven't passed development tests, much less operational tests. Yet they're being hoisted into their silos as we speak—10 in the coming weeks, 10 more scheduled for next year.
According to Graham's Post story, the Pentagon's testing officials estimate in an internal briefing that the interceptors now being deployed would have only a 20 percent chance of shooting down a nuclear missile. An official familiar with the briefing told me that the 20 percent estimate was "on the high side" of a range of probabilities. (The MDA calculated that each interceptor would have an 80 percent chance of hitting its target, but few outside its headquarters seem to take that estimate seriously.)
It's worth noting that these interceptors are among the simplest elements of the missile-defense plan. The full system is envisioned to entail three defensive layers. There is a "boost-phase intercept"—satellite-based weapons, orbiting over suspected nuclear sites, that would shoot down enemy missiles minutes or seconds after they're launched, preferably while they're still ascending in the atmosphere. Then there is the "midcourse intercept"—missiles or lasers on airplanes and ships, as well as in silos on the ground, that would shoot down the missiles as they arc through their trajectory in outer space. (The Fort Greeley interceptors are designed for midcourse intercept.) Finally, there is the "terminal intercept"—missiles, mainly modified versions of Patriot air-defense missiles, that would shoot down the attacker's nuclear warheads as they plunge to their targets in the United States.
For the moment, the more exotic weapons in this scheme are years away from being developed, much less built. The un-modified Patriots, as we saw in last year's war in Iraq, still have trouble distinguishing missiles from airplanes (and have never successfully tracked multiple missiles simultaneously). Finally, all of these elements would have to be linked in an automatic data-processing network more complex than any such network ever constructed. The network would link a) the early warning radar that detects the launching of missiles to b) the release codes that fire the first layer of missile-defense weapons to c) the sensors that guide those weapons to the target to d) the sensors that detect whether the target was hit to e) the release codes that fire the next layer of missile-defense weapons … and on it goes, all the way to the final layer. The entire network would have to operate automatically (there would be no time for human control), and it would have to work without ever having been truly completely tested. Commercial software systems, such as Windows, go through several debugging cycles even after they've been released to the marketplace, i.e., after the engineers back at Microsoft think they've fixed all the problems. This is a normal process; everyone expects it. With the software system for missile defense, there would be no chance for debugging. It has to work the first time out.
According to some officials, the engineers working in the bowels of the missile-defense program know this. They are proceeding as if it were a normal research and development project. They know that, at this stage, technically, that's all it can be. The thing is, President Bush and Secretary Rumsfeld see it as a fast-track procurement program ready for—and undergoing—deployment. The president has now heard, from Gen. Cartwright, that this expectation is unreal. This may be why Bush has refrained from proclaiming another "mission accomplished." The question now is whether he'll take the next step and realize that, given the program's true status, there's no point lavishing it with so many billions of dollars, especially when "homeland defense" against more real and urgent threats are barely getting millions.