“Chatter” seems too casual a word for what is arguably the most important single product of the mammoth American cyber-industrial establishment which gathers “communications intelligence,” commonly abbreviated as Comint. Intelligence professionals use “chatter” to describe the miscellany they acquire of the personal and operational communications of “persons of interest,” another term of art meaning people who may know or be planning something the United States wants or needs to know about. For the last three years the people at the top of the American list of persons of interest have included Osama bin Laden, his lieutenants, associates, and supporters in al-Qaeda, and the widening circles of Islamic fundamentalists who share or know or have heard rumors about Osama’s goals and plans. In the absence of agents reporting from al-Qaeda’s innermost sanctum, American intelligence professionals must depend on chatter to keep track of whatever devastating attacks al-Qaeda’s terrorist cells may be planning next.
Over the spring and summer of 2001 intercepts of terrorist chatter rose to dramatic levels but were shrugged off by the White House and the President’s then national security adviser, Condoleezza Rice. The destruction of the World Trade Center ended that; now the nation’s electronic ears strain for every terrorist whisper. Periodic official warnings of new attacks on tunnels and bridges, on major sports events, on commercial airliners arriving from France, and on the New York City subway system have all been identified as prompted by chatter, often described as reaching levels not seen since just before September 11—shorthand for “listen up, this is serious.” In the case of New York’s subways some 16,000 law enforcement per-sonnel in and out of uniform were mobilized after Comint analysts lifted a single worrying word from the chatter—“underground.” What did it mean? No one knew, but responsible officials were not about to wait and see.
But what is chatter exactly? As an American graduate student in Britain in the late 1990s Patrick Keefe came to the subject through news stories about “Echelon”—the code name, first published in 1988, for a coordinated, decades-long global effort by English-speaking countries to intercept communications of intelligence interest. The reluctance of governments to explain what they were doing and why encouraged periodic waves of popular fear that around the corner loomed the omnipresent Big Brother of George Orwell’s novel 1984. These alarms were given semi-official voice in 1998 after a committee of the European Parliament, concerned that Echelon was aimed at them, commissioned a social scientist to write “An Appraisal of the Technologies of Political Control.” The report’s author stated flatly that the eavesdroppers did not just have the technology to listen in; they used it:
Within Europe, all email, telephone and fax communications are routinely intercepted by the United States National Security Agency, transferring all target information from the European mainland via the Strategic hub of London then by Satellite to Fort Meade in Maryland via the crucial hub at Menwith Hill in the North York Moors of the UK.
That word “all” was intended to attract attention and it did. The cold war had been over for a decade, and terrorism had not yet replaced the Red Army as a threat; what were the listeners listening to?
Keefe began his investigation in the pre–September 11 days, before al-Qaeda proved that mega-terrorism was not an idle threat. Things look very differently now when it is widely believed that a few more listeners, quicker off the mark, might have prevented the catastrophic attack that brought down the World Trade Center. Just how close American intelligence came to acting in time, and the deeply rooted reasons it did not, can be found in Blind Spot, Timothy Naftali’s useful new history of the American education in counterterrorism since the Second World War.
That education began with overexcited (and unfounded) fears of a die-hard Nazi terror campaign in the last days of the war, then adapted to the back-alley violence of the early days of the cold war, when the United States and its allies supported resistance groups in the Ukraine and the Baltic countries, and the KGB’s Department 13 responded with attempts, some successful, to assassinate anti-Communist activists plotting against the Soviets.
But it was not until the Six-Day War in 1967 that Americans really began to grasp the challenge of terrorism. In the wake of Israel’s whirlwind victory, Palestinians fought back with terror, the traditional weapon of the weak, and the early cross-border attacks on Israel, rarely successful, were soon followed by attacks in a larger arena harder to defend—pretty much the whole of Europe. Bloody assaults of the sort that targeted travelers lined up at ticket counters never achieved much, but some governments, like the mullahs’ Iran and Qaddhafi’s Libya, grew confident over time that they could attack America and its allies at will with little fear of retaliation.
Naftali is a scholar of national security issues who has specialized in secret intelligence, and Blind Spot is built on long familiarity with the secret organizations responsible for the “counter” in counterterrorism. Few Americans will recall much beyond the names of once-feared terrorists like Abu Nidal and Carlos (“the Jackal”) for the good reason that both were driven from the field by an aggressive CIA-supported campaign to cut off their funds, arrest or kill their operatives, and warn their government sponsors to stand down or accept the consequences. But paradoxically, Naftali argues, those successes did not take. Vigorous counterterror efforts were followed by growing White House timidity, which contributed to the now much-discussed risk-averse culture of the CIA.
The ordeal of the Iran-contra scandal was partly to blame for driving counterterror efforts into reverse, and the rest, Naftali writes, can be blamed on “bureaucratic languor”—his phrase for the go-slow, think-twice institutional caution which follows when presidents lose their stomach for the hard decisions required to fight dirty wars. Naftali does not make the argument that getting tough is all it will take to win the war on terror, but he points firmly to a fatal White House omission during the two or three years before September 11—recognition that al-Qaeda was capable, determined, and dangerous.
But the new American determination in the war on terror does not provide a kind of blank-check justification of whatever it is that intelligence organizations might like to do. The Echelon which caught Patrick Keefe’s interest in the late 1990s may have added some Arabic-language speakers to its staff, and the mood at various communications intelligence (Comint) headquarters may have a new urgency, but Echelon itself remains the same. Official explanations of what that urgency may involve do not proceed from the general to the specific. The phrase “matters of intelligence interest” is about as concrete as any announcement gets. Wanting more, Keefe set out to answer this question for himself, and his book, Chatter: Dispatches from the Secret World of Global Eavesdropping, recounts his travels and conversations with many people worried about government intrusion and a few people with actual experience of the Comint world.
The result is a kind of naturalist’s ramble around the fenced perimeter of the whole vast establishment of technical gear used for intercepting communications. Some things become abundantly clear. For example, it’s called “global eavesdropping” because listening posts circle the globe. The big Echelon stations, Keefe tells us, are Morwenstow in Britain, Sugar Grove in West Virginia, and the Yakima Training Center in Washington State. Secondary stations come and go with changing politics and borders; once-important listening posts in northern Iran, Hong Kong, and southern Germany have been abandoned, while others have been cranked up to take their place.
Some of these date back to the earliest days of British-American cooperation in collecting Comint, like the nine-hundred-person Composite Signals Organization based on Ascension Island in the South Atlantic, where green sea turtles are welcomed on their annual migration from the coast of Brazil but no one else may step ashore without official clearance. The rules are the same on Diego Garcia in the Indian Ocean, retained as a territory by Britain when it granted independence to the Seychelles and Mauritius in the 1960s. To simplify the secret-keeping, the British simply removed Diego Garcia’s two thousand people to Mauritius, where they soon began to pine for home. After decades of supplication a visit was planned for November 2001, but the war on terror intervened. The island is now too busy as a listening post and air force base to tolerate visitors who were, in point of fact, born there.
Chatter contains a lot of information of this sort—names of bases and organizations, descriptions of technologies used for collecting information, a sketchy outline history of the Anglophone alliance beginning in 1946 with the original British and American agreement. Over time Canada, New Zealand, and Australia joined the club but the Americans remain dominant, putting up the money and doling out the take. Other countries maintain listening programs of their own but none can match the American reach, a fact that worries Europeans. In the course of his travels Keefe met just about everybody trying to penetrate the secrets of Echelon and gradually he learned what it is—up to a point.
His education began with the remark of a former British intelligence officer, Alistair Harley, who told him that “if it uses radio waves then [it can] be intercepted, monitored, stored.” The “it” here means people trying to communicate, and radio waves are used to convey just about everything that isn’t written down on hand-delivered paper or transmitted over secure landlines. The most important single fact about Comint is that there is a lot of it, roughly all the daily communications of the politically or commercially active part of the human race, both written and spoken. How much talk are we talking about? Well, it certainly exceeds the conversations (if we could overhear and record them) of everybody in a baseball stadium during a World Series game, plus the conversations of everybody watching it on television, plus the remarks that all the watchers combined have made or heard over the last…week? year? decade? There is a lot of talk going on, and all the talkers are talking at once. The problem is to separate the words of urgent interest from the roar of background chatter. This is where Echelon comes in.
“Echelon,” Keefe learned, “is nothing more than a secret code name for a specific computer program used to sort through intercepted satellite communications.” Lest reference to satellite communications be interpreted as bringing the problem down to size, Keefe reminds us that over the course of the 1990s the number of cell phones in use increased from 16 million to 741 million, the number of Internet users increased from 4 million to 361 million, and the annual number of minutes devoted to international telephone conversations increased from 38 billion to 100 billion. At one point or another almost all of that talk passes from the surface of the earth to a satellite and back again—the moment when the eavesdroppers pick it up. But the sorting job is beyond human eavesdroppers; half the population would be required to listen to the other half. “Echelon,” Keefe writes in a second pass at definition, “refers to a particular type of computer that is used to sort through large amounts of data for items on a given watch list.” Keefe has not learned much about exactly how this is done, but even that little is helpful in providing a sense of the magnitude of the challenge.
The work, Keefe explains, begins with establishing two lists—one of persons and organizations (the “watch list”), and a second of “keywords” that make the intelligence ear perk up. These lists are collected in “Echelon Dictionaries,” which are maintained by dictionary managers responsible for putting names and words in or taking them out. Osama would be on the watch list, and “bomb” and “attack” and “anthrax” and “New York City’s water supply” would be among the keywords, along with many others. Finding them in the ocean of chatter begins with “packet sniffers,” which check the “data packets” that transmit communications in electronic form. When the sniffer finds a match between address or message and watch list or keyword, it copies the message for further examination—the moment at which human beings begin to be involved.
Here written communications present one kind of a problem and voice communications another, generally considered more difficult. To sort out the second type requires a “word-spotting” capacity—essentially a computer program that can distinguish between spoken words in a multitude of languages and is not fooled by synonyms—“device” for bomb, say, or “sniffles” for smallpox. Admiral Bobby Ray Inman, a longtime director of the American codebreaking and Comint efforts as chief of the NSA, once admitted in public that word spotting for voice systems remained a dream. “I have wasted more US taxpayers’ dollars trying to do that,” he said, “than [on] anything else in my intelligence career.”
Classic chatter is the target of these global Comint search efforts, and while everything is within reach, it is hard to spot the significant in time to make use of it. On September 10, 2001, Keefe reminds us, two messages in Arabic were intercepted in the course of transmission from Afghanistan to Saudi Arabia. Both rang the al-Qaeda bell and were retrieved for translation and analysis. When they were read two days later one said, “Tomorrow is zero hour,” and the other, “The match begins tomorrow.” Would translation of those two messages in good time have served as the two-by-four once used by sharecropping farmers to catch the attention of balky mules? No one really thinks so, although vast sums have been recommended and will doubtless be spent to speed up the entire effort to sift terrorist chatter for timely warning of the next attack.
After years of studying this problem, Keefe, like the European Parliament, concludes that Echelon is not intended as Big Brother’s foot in the door, but represents a good-faith effort to solve a genuine problem with state-of-the-art computers and programs which (it is hoped) might succeed in penetrating terrorist cells where traditional spies have been dismissed as bound to fail. The secrecy that envelops Echelon is not sinister but practical; when people don’t know what you’re doing they can’t stop you, and they can’t protect against it.
But chatter is not the only target of the NSA; the listeners also target individuals, groups, and places to monitor conversations that might give the United States secret advantage in the effort to do what it wants to do. The most interesting section of Keefe’s book retells the story, only sketchily reported in The New York Times and The Washington Post as it unfolded, of the joint British and American effort to monitor the communications of UN Secretary General Kofi Annan and the diplomatic traffic of Security Council members as they sought guidance early in 2003 from their home governments for dealing with the relentless American effort to browbeat the council into voting for war on Iraq.
The episode surfaced in public in a manner of utmost rarity—a deliberate leak to the press by a career employee of an intelligence organization, in this case a translator for Britain’s Government Communications Headquarters (GCHQ) in Cheltenham. After four days of agonized indecision the translator, Katharine Gunn, gave a document to a friend with connections in the mass media. A month later it appeared on the front page of The Observer, giving the world a raw glimpse into the boiler-room reality of communications intelligence. The document was a January 31 memo to GCHQ from Frank Koza, chief of staff for regional targets of the American National Security Agency. Koza wrote to his GCHQ colleagues:
As you’ve likely heard by now, the Agency is mounting a surge particularly directed at the UN Security Council (UNSC) members (minus US and GBR of course) for insights as to how to [sic] membership is reacting to the ongoing debate RE: Iraq…the whole gamut of information that could give US policymakers an edge in obtaining results favorable to US goals or to head off surprises.
The surge being mounted was in efforts to bug conversations and intercept communications among UN delegates as part of the American strong- arm effort to get a favorable Security Council vote, but by the time Koza’s memo was published the debate was over and the US was on the verge of going to war. An internal investigation was immediately launched by GCHQ and after a day of denial Gunn told her boss, “The leak was me.”1 A year of cat and mouse followed before the British authorities decided not to prosecute the case, but when Keefe finally arranged an interview with the twenty-eight-year-old Gunn he found that she still was not free to answer even the simplest questions about Echelon:
“But it’s a name you’ve heard before?” I pushed.
“Yes,” Katharine said slowly, looking me in the eyes. “But I can’t comment on that either.”
The only remarkable thing about the Koza memo is that it got published. In every other way it appears to have been a routine exchange between the NSA and GCHQ of the sort that has occurred daily for fifty years. Chatter is the latest in a series of books to explain what Comint is and does, and while it is written with fluid grace and disciplined structure, in truth it does not add much hard new information to what has already been published by the two acknowledged masters of the field—David Kahn, whose history The Codebreakers remains the standard text nearly forty years after it was first published, and James Bamford, whose two books on the NSA have revealed pretty much all the public knows about the black cube in the Maryland countryside once called “No Such Agency.”2
Over the years there have been occasional moments of aroused public scrutiny when intelligence collection is suddenly glimpsed at work—for example, when an American ELINT plane was shot down after ignoring Chinese warnings to back off, or when the CIA’s then chief William Casey threatened to arrest the Washington Post reporter Bob Woodward for spilling the beans about a US program to tap Soviet underseas telephone cables, or when the reporter Seymour Hersh revealed that CIA operatives had been routinely opening first-class mail sent to or by US citizens on a watch list of antiwar activists, or when the US pushed a boatload of NSA eavesdroppers so close to the scene of fighting during the Six-Day War that Israeli planes sank the craft with the loss of thirty-seven American lives. But with those exceptions and a few others, all forgotten with amazing rapidity, the public record is empty of any kind of sustained discussion of what Comint is for or costs, where it succeeds or fails, what it collects, who gets to read it, and who decides. About all the public really knows is that the NSA is big.
It wasn’t always big. In the spring of 1919, when the father of American cryptography, Herbert O. Yardley, drew up a plan for a permanent State Department codebreaking organization—a “black chamber,” in traditional European parlance—he estimated that a modest $100,000 a year would buy a chief (Yardley) and fifty clerks and cryptanalysts. In the event the staff was half that and the budget for the new organization was split by State and Army, which wanted a hand in. Yardley rented a three-story building in New York City, a choice mandated by federal law, which prohibited the State Department from adding new employees to its staff in Washington. There, on East 38th Street just off Fifth Avenue, Yardley put two dozen people to work under civilian cover—as the Code Compiling Company, incorporated in New York to produce and sell commercial codes. Thus concealed, Yardley’s shop was to collect and decipher foreign communications of interest to American diplomats and soldiers.
The first challenge was to get copies of cables, in theory protected by law from release to unauthorized persons. Yardley got around this easily enough by simply asking cable companies to hand them over. What he told them is unknown, but it worked. The head of the American Cable Company told him, “The government can have anything it wants,” and other firms were also obliging in varying degree. By the end of its first year the Cipher Bureau, as it was formally known, was solving and reading a multitude of foreign codes.
Yardley is one of the remarkable men in American history. He is known primarily for his summary dismissal in 1929 by incoming Secretary of State Henry Stimson, a patrician Wall Street lawyer who closed down the Ci-pher Bureau with the casual observation that “gentlemen do not read each other’s mail”—a remark, interestingly, which is the only thing remembered about either man. It is often cited as marking the high-water mark of American starched-collar idealism before the downhill slide into great-power realism. But what made Yardley famous is not the thing that makes him interesting. The son of a railroad telegrapher, a man with a lively Jazz Age interest in money, good-looking women, and drinks at five, Yardley not only taught his country how to read other people’s mail but wrote two of the enduring American books—the best single intelligence memoir, The American Black Chamber (1931), and perhaps the greatest book in any language on playing cards for money, The Education of a Poker Player (1957).
Either might have justified a biography long since but two barriers intervened. One was the enduring anger aroused among officials by Yardley’s cool account in the first of his great books of the routine deceptions of government when no one is watching. Fair judges might have noted that Yardley was out of a job and needed the money when he wrote The American Black Chamber, and that the government had announced in effect that it was getting out of the codebreaking business for good, leaving Yardley’s revelations interesting but moot. Such excuses cut no ice; in the three decades that Yardley survived the Cipher Bureau, official anger never relented a degree. Time and again Yardley was blocked, ignored, blacklisted, or secretly undermined—an eloquent demonstration of what intelligence professionals and their employers do to those who break the code. Laws against speaking out of school are hardly necessary; the freeze-out is enough.
But Yardley remains the great figure of American codebreaking and it was probably inevitable that David Kahn, the great historian of American codebreaking, would set out to write his biography. From the outset he was challenged by the second major barrier to writing Yardley’s life—lack of materials. When Yardley speaks in his books—there is a third covering his adventures in China in the 1930s—all is illuminated, but where the books stop the life grows dim. At his death, Yardley left no papers—odd for a writer—but when intelligence figures die it is not uncommon for personable men to arrive promptly at the widow’s door with an offer to help. Typically the visit ends with every scrap of paper going out the door before the sun goes down. Kahn offers no guess about the fate of Yardley’s missing papers, and repaired the deficiency in the only way—by scouring every plausible archive, talking to the bare handful of survivors, and trusting to luck. His big finds were the files of Yardley’s literary agent, George Bye, preserved in the library of Columbia University, and the letters Yardley sent home from China during the year and a half he worked for the legendary chief of Kuomintang intelligence, Dai Li.
The man who emerges in Kahn’s briskly paced portrait is gifted, complex, resourceful, and often disappointed. Yardley’s life included more periods of drinking than not, some interesting women, and many spurned efforts to resume the work he knew and liked best. He bounced back from the loss of his codebreaking job with The American Black Chamber, hung around Hollywood long enough to earn $10,000 for doing nothing, wrote some forgettable novels, did some radio work, dabbled in real estate, and finally got back into the great game, attacking Japanese codes for officials in China. During after-hours in the Chungking Hostel he taught the young reporter Theodore White two useful survival arts—how to play poker and how to ride out an air raid, later summarized by White in a memoir:
The chief danger of an air raid, he said, was splintered glass from windows. Thus, when one hears the siren one should get a drink, lie down on a couch and put two pillows over oneself—one pillow over the eyes and the other over the groin…if the eyes or groin were injured, life was not worth living. It was good advice for any groundling in the age before atom bombs; and I took it.
After leaving China in mid-1940 Yardley briefly worked for Canadian intelligence, but a few sharp words from Washington ended the job and seemed to take the fight out of him. Back home he was employed by the Office of Price Administration for the last half of World War II, then moved sideways in 1947 to become a public housing bureaucrat. He died in 1958.
We might study Yardley’s life with profit for the poker, but Kahn of course gives pride of place to the making and breaking of codes, and there are wonderful accounts in The Reader of Gentlemen’s Mail of some of Yardley’s greatest feats. Here Kahn’s mastery of the field gives his book genuine intellectual excitement; solving a code involves a kind of personal combat like chess, with the added drama that the loser generally does not learn of his defeat until it is too late.
Kahn is frank to admit that Yardley was not the greatest American codebreaker; his chief rival, William Friedman, was a duller man, but he broke many wartime Japanese codes, was smarter than Yardley, did more to advance the art, and solved wartime messages that made a bigger difference to the course of history. But Yardley was first on the scene, he left an account of the secret world which is unsurpassed, and on one notable occasion his solution of a code allowed US diplomats to sit tight in a down-to-the wire negotiation limiting Japan’s navy because they had read the official instructions telling Japan when to give in.
The scene was Washington, November and December 1921. The world’s naval powers had come to negotiate limits to shipbuilding to prevent a runaway naval race and save money. The point in contention was the ratio of tonnage afloat between the three largest navies, those of Britain, the United States, and Japan. The US proposed a ratio of 10:10:6—limiting the British and Americans to 500,000 gross tons of naval shipping, and the Japanese to 300,000 tons. But the Japanese were unhappy and would not budge from their insistence on a 10:10:7 ratio, which would give them 350,000 tons. As Kahn makes clear this was not a distinction without a difference. Calculations difficult to summarize here meant that Western navies would be at a disadvantage in Japanese waters with a 10:10:7 ratio, but would have ships enough to dominate even far from home ports if they could insist successfully on 10:10:6. It was at this moment that Yardley earned his place in history.
Two years earlier after months of work Yardley had solved an important Japanese diplomatic code; a later variant was broken in the summer of 1921, and on December 2, as the naval conference struggled over its impasse on the ratio, a copy of a cable from Tokyo was delivered to Yardley’s team and deciphered almost as quickly as a clerk could type. The drift of the message, contained in sixty-three ten-letter groups, was an instruction to Japan’s negotiators to defend the ratio tenaciously, falling back one by one through the four positions only as required to prevent the negotiations from breaking down entirely. As Yardley later described in The American Black Chamber, infuriating official Washington, position number four was agreement to the 10:10:6 ratio. “Stud poker,” Yardley wrote, “is not a very difficult game after you see your opponent’s hole card.” So it proved. On December 12 the Japanese caved. Set aside all the other solutions achieved by the Cipher Bureau during its ten-year life for a total expenditure of only a third of a million dollars; “this alone,” Kahn writes, “made it worth the money spent on it.”
But as so often in life and in the intelligence business, things were not entirely as they seemed. Yardley was not the only one trying to read Japanese intentions. A new Japanese ruler, Crown Prince Hirohito, had just taken power in Japan. On November 28, four days before Yardley even saw the critical cable to Japan’s chief negotiator, The New York Times reported that “Tokyo Is Prepared to Yield on Ratio.” Two days later it told readers that the crisis had passed and agreement on the stricter ratio was likely in a week’s time. How did the Times achieve this feat of prediction while Yardley’s Cipher Bureau was still decoding old messages telling the negotiators to stand fast? The short answer appears to be reporting; they stuck close to the Japanese negotiators, and when the wind began to change, one of them told the Times what to expect.
Kahn is of two minds about this turn of events. As a lover of neatly wrought code solutions he is full of admiration for what Yardley did and the confidence it gave American diplomats to wait patiently for the apple to fall in Washington’s lap. Codebreaking, he writes at the end of the book, “alone provides believable, high-level, unmediated, voluminous, continuous, cheap information.” True enough, when the analysts find the solution and find it in time. But as a realist Kahn also admits that there was something anticlimactic about Yardley’s greatest achievement. He helped his country win a hand in high-stakes international poker, and was rewarded for his feat with a special commendation and a bonus of $184, but his solution changed nothing: that hand was going to the Americans anyway.
Has the globe-encircling NSA done better than Yardley’s tiny bucket shop on East 38th Street? Have the billions spent on satellite collection systems and computer programs like Echelon delivered value for money? Have they made America safer? Intelligence professionals whisper about seldom-touted successes and Patrick Keefe concedes in effect that even a blind hog will find the occasional acorn. But his final judgment is harsh: “Chatter is, as it turns out, a perfect word for the conversations culled from the airwaves: fickle, misleading, most often inconsequential.” September 11 was the test. No matter how success is defined by the intelligence world, the Anglophone countries and their listening posts fell short. Comint, Keefe states bluntly, “had its day and failed.”
About the failure everyone now agrees. But what was the problem? And what should be done to make us safe? Keefe has no idea. Sounding a little dispirited after his years of research and writing, he urges Americans to think hard about where to draw the line between liberty and security, but it’s an odd note on which to conclude. It wasn’t respect for the Constitution that kept the NSA from reading the “Tomorrow is zero hour” message until the day after the disaster. It was lack of translators. To meet that kind of problem, the Comint professionals have a default solution: more. Not just more Arab linguists but more of everything—more analysts, more polygraph examiners and security guards, more freedom to listen in on more people, more listening posts, more coverage, more secrecy. Is more what we really need? In my opinion not. Ordinary reporters scooped Yardley in 1921, and ordinary spies—human agents, run by case officers in the field—are most likely to penetrate the heart of terrorist circles now. But running spies is not the NSA’s job. Listening is, and more listening is what the NSA knows how to organize, more is what Congress is ready to support and fund, more is what the President wants, and more is what we are going to get.
May 12, 2005
See Patrick Radden Keefe, “‘The Leak Was Me,'” The New York Review, June 10, 2004. ↩
David Kahn, The Codebreakers: The Story of Secret Writing (Macmillan, 1967); and James Bamford, The Puzzle Palace: A Report of America’s Most Secret Agency (Houghton Mifflin, 1982) and Body of Secrets: Anatomy of the Ultra-secret National Security Agency from the Cold War Through the Dawn of a New Century (Doubleday, 2001). ↩