I don't believe that the U.S. plane was in international waters. I don't believe that the U.S. expects the Chinese to honor demands to return the plane untouched. (In 1976, when a defecting Russian pilot flew his MIG-25 to Japan, the Soviet Union demanded its plane back. We eventually returned it, in parts.) And I know this kind of thing is business as usual.

The United States has flown spy missions over other countries since the 1950s: the Soviet Union, China, North Korea, Vietnam, Iraq. The Soviet Union used to fly them over the United States. The target country would routinely launch fighters to harass the spy planes. This was where the Cold War would get warm, as the pilots buzzed each other, called each other names over the radio and made obscene gestures out their windows. Not all of these flights ended well. In 1956, the Chinese shot down an American spy plane in the East China Sea, killing 16. In 1968, the North Koreans shot down another spy plane, killing 31. And in 1960, the Russians downed Francis Gary Powers' U-2 spy plane as it passed over their country. Inside the NSA's B2 Operations Building, there is a monument of black granite with the words "They Served in Silence" and the names of the 152 military and civilian eavesdroppers who have died, most of them on ships and planes, peeking up the electronic skirts of our adversaries.

Bamford's closing chapters are cautionary ones. Today the NSA is being flooded by a fire hose of communications, while at the same time it is being denied other communications through never-ending improvements in communications technologies. Satellites are trivial to eavesdrop on; fiber-optic cables are very difficult. The Internet has its own challenges. But most of all, the NSA's problems lie in the difficulty of interpreting intelligence, not in the difficulty of collecting it. I have long believed that the NSA's future lies not in intercepting communications but in targeting static databases: data at rest as opposed to data in motion. Bamford agrees.

All this makes the China incident even more confusing. I don't understand why, in a world where intelligence satellites can eavesdrop on anything anywhere, where ground stations in Japan and South Korea have China well covered and where massive intercept programs like Echelon vacuum up almost all foreign telecommunications, we need to launch aggressive and provocative spy missions against countries like China. I can't think of another midair collision that didn't end up in two crashed planes; it's a miracle that the American EP-3 survived. If the 24 Americans had died as a result of this incident, how would Congress have reacted? Would we have believed China's claims that it was an accident, not an attack? Would we have so easily turned our warships around after the Chinese government refused our offers to assist in recovering the wreckage? How much more aggressive would the rhetoric have been on both sides? I don't mean to imply that the U.S. deliberately set out to cause an international incident, but it seems to me that it was ignoring some pretty obvious risks for some pretty dubious rewards.

Fortunately, the plane's crew members weren't killed, and we didn't have to face the kind of crisis their deaths would have triggered. But Bamford's book explains the secret history of times when the rhetoric was more aggressive, when enemies would shoot each other down and when what the world's leaders said in public did not match what they did in private. It's a sobering history, and one we should take pains not to repeat.

Maya Angelou reads from "The Heart of a Woman"