We hardly ever get much of a peek into what our $28 billion- a-year spy agencies really know. Recently though, a source provided me with a bibliography of secret intelligence reports prepared over the past few years by the Defense Intelligence Agency.

The DIA list might not answer the big question, but it offers fascinating insights into the world of the professional paranoids. DIA is not the sexy agency, and its dogged military monitoring is more akin to record-keeping than espionage. And while it is hard to garner much Washington chatter that is complimentary about an organization that sometimes seems stuck in old Cold War ways, one thing is clear: The DIA is busy.

DIA is the military's diminutive sibling in an intelligence triumvirate with the CIA and NSA. Established in 1961 to consolidate military intelligence, the Washington-based agency has evolved as a broker of Defense Department collection requirements.

Though DIA has a force of military attaches at embassies worldwide and supervises the small Defense Humint (Human Intelligence) Service of case officers and foreign agents, it is mostly an analytic agency.

DIA, in the words of a Pentagon briefing, manages "intelligence production." If this term has an unfortunate industrial overtone, it is because there is a certain mechanistic tenor to its work.

With its various affiliates - the national air, ground, maritime and medical intelligence agencies - DIA maintains the "baseline" intelligence military threat assessment, which is essentially the knowledge base that resides in a series of humongous databases of weapons and targets.

What does the workforce, which approaches 8,000 people, do? They sift through incoming raw intercepts, agent reports, photographs and translations, extracting what is of military importance, writing time sensitive dispatches, updating their databases and producing "finished" reports.

Technical Perfection

To be clear, according to the 1993-1999 bibliography of reports labled "Secret," a good 80-plus percent of production deals with the former Soviet Union, China, North Korea, Iraq and other rogue states. There are intriguing titles like "Kenyan Relations with China and Iran" (1997) and "Havana: VIP Emergency and Routine Medical Treatment" (1993) in the mix, suggesting insights and prognostications beyond what is in the news. But by and large, most of this output is the extremely detailed and mostly unglamorous effort that goes into building the "order of battle" of military forces and helping to determine the true technical characteristics of weapons.

The breakup of the Soviet Union and the fall of the Warsaw Pact has meant a bonanza for technical intelligence. All around the globe, foreign gear is secretly acquired by the United States, and "foreign material exploitation" studies are done to develop U.S. weapons and countermeasures. Some 150 exploitation reports are produced annually, according to the bibliography, all with code names like "Heart Ache," "Sensor Flower" and "Ibis Trough" that obscure the identity of the equipment being tested.

But it is not just former Soviet equipment that is of interest. Given that a French airplane is just as likely to show up in an adversaries's arsenal (as they did in Iraq in 1991), secret assessments of friendly equipment are also required (such as the 1998 study of the French Barax/PAJ-95 defensive jammer). And increasingly there is high-level interest in foreign information processing and telecommunications networks and microelectronics. The first "Cellular Communications Systems Worldwide: Technical Overview" was completed in 1994. More than two dozen technical reports on foreign networks were done in 1998 (undoubtedly even more are produced at high levels of classification than "Secret").

Another series of intelligence assessments is done under the title "weapons acquisition strategy." Given increasingly diverse domestic military production and the competition between arms suppliers, it is no wonder that countries provided in the past two years include Sweden, South Africa, Japan, Mexico, Brazil, Switzerland and Saudi Arabia.

A Wide-Angle Lens

There is no doubt that American technical intelligence - understanding foreign weapons and building countermeasures - is a strong point, as well as an effort that can be directly attributable to zero combat casualties in Yugoslavia. Yet intelligence requirements in the post-Cold War era are slowly shifting from focused scientific and technical intelligence to economic, sociopolitical and paramilitary issues.

The top-secret Presidential Decision Directive 35 ("Intelligence Requirements"), signed by President Clinton in March 1995 (see box), articulates these post-Cold War needs. DIA's core mission, meeting "the intelligence needs of our military" is still the highest national priority. But now, in addition, high priority is assigned to "trans-national threats to our security, such as weapons proliferation, terrorism, drug trafficking, organized crime, illicit trade practices and environmental issues of great gravity."

PDD-35 also establishes tiers from 0 to 4 defining intelligence requirements. Tier 0, of greatest urgency, is "crisis coverage." Tier 4 are countries virtually of no interest to the United States. The suggestion, of course, is that there is some area of the globe which doesn't need to go under the microscope. Yet when all the demands are met, trans-national issues covered, Tier 1a hot topics folded in, no country remains where the president explicitly says there should not be monitored. The bibliography proves that, in spades.

There are "military capabilities studies" not just for the obvious Libyas and Serbias, but also for Jamaica, Seychelles, Swaziland, Bangladesh, Japan, Nepal. Orders of battle are compiled not just for Iraq and Iran, but also for NATO allies, for Australia/New Zealand and Singapore. Secret "U.S. only" reports are written about "Guatemala's Corps of Engineers" (1996) and "Kuwait: The Future Role for Women in the Armed Forces" (1995), and other obscure topics of tentative vital importance.

If Everything Is Strategic, Nothing Is

As has been demonstrated by operations in Somalia, Haiti, Rwanda, Sudan or East Timor, no corner of the globe can be too obscure for U.S. intelligence to monitor. Someday, someone will be called upon to be a Rwandan military expert. Of that, the intelligence community seems secure.

But to what end? Does all this analysis help decision-makers to set priorities, to anticipate dangers, or does the broad-based record keeping make military operations smoother and more prone to success? The answer is an unequivocal no.

Part of the problem is the inability to stop what has always been done, to admit that the end result of intelligence analysis is not a better or more intelligent government. Take for instance, medical intelligence, which failed to predict or detect the Gulf War syndrome and seems bewildered by widespread soldier and veteran despair when it comes to health care. A mind-boggling array of "medical capabilities studies," urban medical profiles and "health services assessments" are prepared.

No doubt "Facing Winter: Public Health Conditions in Kosovo from November 1998 through March 1999" was a timely report. But what about the secret medical study of the Caribbean and Mid-Atlantic islands completed in 1997, or other recent reports on Burkina Faso, Sao Tome and Principe, Chile or Austria? Surely "High Risk Snakes of the Middle East" is an important evaluation for Army doctors and troops to have, but does it have to be classified "Secret," and do government intelligence analysts need to do it?

E-Mail and Tapped Phones?

One result of covering the globe with some level of detail is that the DIA and its big brothers often miss the big events, such as an Indian nuclear test (too few analysts assigned, said the government's assessment of the bungle). That old adage that a little knowledge is a dangerous thing really is true. Equally so, growing Internet mania about government monitoring of E-mails and phone calls demonstrates a lack of appreciation of just how pedestrian and uninteresting much intelligence work is.

Since PPD-35 evidently seems incapable of refocusing the intelligence agencies away from the marginal, it is not surprising that skeptics would be right to dismiss its language which identifies "targets that the U.S. intelligence community will not collect against," such as U.S. citizens who are not of foreign or counterintelligence interest. Still, with dozens of databases to fill and update and literally thousands of reports to prepare on every subject and country from A from Z, one has to wonder just how many analysts are left to pry into the lives of the fearful.

William M. Arkin can be reached for comment at william_arkin@washingtonpost.com.