Fact and fiction converged for me in a story that changed my life -- and turned me into a novelist. That was a front-page story I wrote for the Wall Street in February 1983 revealing that the CIA had recruited Yasser Arafat's chief of intelligence as an American asset, and had run him off and on until he was assassinated by the Israeli Mossad in 1979.

I worked on the story for more than two years: I began my assignment as the Journal's Middle East Correspondent in 1980 with one piece of information that someone had let slip in a conversation that summer in Washington, as I was preparing to leave: My informant said that the previous year the Israelis had assassinated a person, whom this man called "our man in the PLO." After dropping this astonishing tidbit, he wouldn't say more, but after a little research I realized that he must be referring to a key PLO operative who'd been killed in 1979, named Ali Hassan Salameh, alias Abu Hassan, who was Arafat's chief of intelligence. To the Israelis he was the "Red Prince," a member of "Black September" and one of the architects of the Munich hostage-taking that led to the slaughter of Israeli athletes. This was a story that seemed to have everything -- but first I had to find out if it was true.

In Beirut, which was my most frequent address in those days, I sought out people I thought might be able to shed light on this extraordinary tip. The Lebanese civil war was still going on, and I spent as many afternoons as I could in the company of one Palestinian who I was certain must know the story. He pointed me toward a Lebanese man who knew all the details, and I eventually tracked him down. Finally had enough material to publish a front-page story. It opened with the scene in which President Carter is informed by his CIA chief, Stansfield Turner, about the assassination of this man. The story quoted several former U.S. officials, on the record, who confirmed Salameh's assistance to United States and the lives that had been saved by this secret relationship with Arafat's chief of intelligence. I knew at the time that the CIA officer who had run this operation was named Robert Ames. I did not identify him in the story because of the risk that he might be killed if his identity was disclosed.

Tragedy turned this story very dark: On April 18, 1983, a terrorist truck bomb destroyed the American Embassy in Beirut. On a freak chance, Robert Ames happened to be visiting the embassy that day on a trip from Washington, where he was now serving as National Intelligence Officer for the Middle East. Ames was killed in the bombing, along with every other member of the CIA station who was in Beirut that day.

I was inside the embassy on the day of the bombing. I am interviewing a military attache. I left at about 12:30. I go back to my hotel. Just after 1:00 pm the car bomb exploded. The bomb blast was the loudest I had ever heard in Beirut. I ran back down the hill from my hotel to the embassy site. It was already ringed with Marines. Late that day, one of my Lebanese sources who had been planning to have dinner that night with Ames said that he was missing. The next day, the State Department confirmed that he was dead.

In the aftermath of Ames's death, the Arabs who had been working with him and his colleagues, who had deep bonds of attachment with them, needed to grieve. I was the only American left in town who really knew the story, because I'd been working on it for two years. They knew I knew it. And so they sought me out and began to tell me details about intelligence operations -- the wiring diagram details that it would be impossible to publish.

That's when I became a novelist. It was obvious that the only way I could share this world of fact was through fiction. Other, the story was impossible to write: It was too raw, and at that time genuinely dangerous. I wrote many drafts, and the novel was eventually published by WW Norton in 1987 as Agents of Innocence.

When I began serious work on the novel I put a picture on my desk of Robert Ames, this extraordinary case officer who died in the embassy. I had cut it out from the obituary that ran in The Washington Post. and put it in a frame. But after a few weeks, I put the picture in my desk. I realized that what makes a novel seem real, paradoxically, is its departure from actual life: The way it's re-imagined in the mind of the writer gives it a reality and power that it wouldn't have otherwise. Otherwise, you'd just be reading a 120,000-word newspaper story.

I'd never written a novel before, and wasn't sure how to do it technically -- how to tell the story, move the characters around, divide up the slices of time. On the advice of a friend who was both a novelist and journalist, I opened each chapter with a dateline, as in a newspaper story, specifying when and where that scene took place. That helped with composition.

Although that book was sold as a novel, CIA officials and PLO leaders knew it was a real story. On the CIA's web site for many years, the book was listed among the recommended books about operations, with a note that said, "Though a novel, senior officers say this book is not fiction." I am told that at the agency's training facility near Williamsburg, known as "The Farm," the book is often recommended to young recruits. Often when traveling abroad in recent years, I have been approached by agency officers who said that when they wanted to explain to a parent or spouse what the CIA really does overseas, they gave them a copy of Agents of Innocence.

I wrote several more newspaper articles during the late 1980s about the secret CIA relationship with the leading terrorist organization of the day. And after the 9/11 attacks, when everyone began screaming for intelligence from inside Al Qaeda, I wrote this column explaining how the CIA-PLO operation had worked:

Penetrating Terrorist Networks

By David Ignatius

Sunday, September 16, 2001; Page B07

As the United States prepares for war against a shadowy enemy, politicians are understandably insisting that the CIA improve the quality of its information. America should never again feel as blind and vulnerable as it did last Tuesday.

But some of the second-guessing from Capitol Hill has implied that fixing the CIA will be an easy task: Just add more money and spies and -- presto -- you have better intelligence about terrorist networks. Alas, if spying were that easy even a member of Congress could do it.

To understand the challenge facing America's spymasters in coming months, it's useful to recall the CIA's most successful operation ever against terrorists -- its penetration of the Palestine Liberation Organization during the 1970s. I reported that story during my years as Middle East correspondent for the Wall Street Journal, and later wrote a novel, Agents of Innocence, which described the operation in detail.

What the CIA did back then was to open a secret channel with the PLO's chief of intelligence, a charismatic operative named Ali Hassan Salameh. At that time, Palestinian terrorism against the United States was a serious threat -- as vitriolic if not as deadly as Osama bin Laden is today.

When these secret contacts began in 1970, they had the explicit approval of Yasser Arafat himself. But the relationship evolved and expanded over the years, and Salameh came to be regarded within the CIA as a witting "asset" -- uncontrolled and unpaid but immensely valuable. He supplied the CIA with extensive information about terrorist groups, and on several occasions he intervened directly to halt planned attacks. His cryptonym in CIA cable traffic, I'm told, was MJTRUST/2.

Part of what motivated Salameh to work with the CIA was the belief that America could help the PLO win its political goals. It was America's unique status -- a close ally of Israel but also a peace broker allied with moderates in the Arab world -- that allowed the intelligence relationship to work. Assets as valuable as Salameh can rarely be "bought." The motivation is deeper and more complex.

The CIA officer who handled Salameh was an equally charismatic man named Robert Ames. He loved the Middle East, spoke Arabic fluently and had a gift for the human relationships that are, in the end, what espionage is all about. People like Ames can't be bought either, no matter how much money Congress appropriates. They are successful because they enjoy their work and are given wide latitude to do it effectively -- without lawyers or legislators looking over their shoulders all the time.

Make no mistake: Salameh, the CIA's contact, was a terrorist himself. He had been a key operative in Arafat's covert "Black September" organization, and the Israelis believed he helped plan the massacre of Israeli athletes at the Munich Olympics in 1972. They even targeted Salameh for assassination. To gather information about terrorism, in other words, the CIA was prepared to sup with the devil.

Even after Munich, the CIA worked to maintain its intelligence link to Salameh. Ames sent the Palestinian this message in September 1973, using a simple code for the CIA, the PLO and Israel: "My company is still interested in getting together with Ali's company. The southern company [Israel] has investigated. I've seen a lot of their files, and they know about our contacts."

Hauntingly, both men were eventually killed by car bombs -- Salameh by an Israeli hit team in 1979; Ames by the Iranian-backed terrorists who destroyed the U.S. Embassy in Beirut in April 1983, on a day he happened to be visiting from Washington. But the secret contacts the two men began eventually led to the Oslo peace accords.

Few intelligence operations work out so successfully. But there are some relevant lessons that can be drawn from this long-ago skulduggery. The most obvious is that collecting intelligence about terrorists is a truly dirty business. This world cannot be penetrated without help from members or friends of the terrorist network. Rules that pass muster with lawyers in Washington won't get you very far in Kabul or Islamabad.

A second lesson is that there aren't many quick fixes in the spy world. Relationships of trust take months and sometimes years to develop. America's allies may be better placed to develop these relationships than the CIA. That's part of why friendships with moderate Arab regimes remain so important; they provide the intelligence liaison that the CIA feeds on.

And, finally, the collection of intelligence can't really be separated from America's larger role in the world. That doesn't mean sympathizing with the terrorists or endorsing their demands. But to win a war, you need allies. And if the United States treats all Muslims as the enemy, it isn't likely to have much luck recruiting them as spies.

Paradoxically, these tragic days have probably been an ideal time for the CIA to be recruiting new sources of intelligence about terrorism. The barbaric attacks Tuesday aroused disgust around the world -- not least among civilized Muslims. Some of these disgusted Muslims will surely want to help the United States and its allies put the terrorists out of business.