In a few days the United States armed forces will attempt to discover if it is possible to successfully place about 500 journalists in military units (down to the company level) going into war. This experiment in what the military calls "embedding" entails grafting what amounts to a presidential-campaign-sized press corps onto an army in combat. The question of whether this is going to work, or implode, is a matter of much conversation among the involved parties here.

On Monday, in the lobby area of the Hilton Resorts, where the U.S. and British militaries have established the Coalition Press Information Center, two such parties were discussing, over cigarettes, a particular aspect of concern. Both were veterans of military-media relations: Max Blumenfeld, a major in the Army Reserve who is the chief of plans and operations for public affairs in the V Corps, and Alisha Ryu, who is a Nairobi-based correspondent with Voice of America.

Blumenfeld and Ryu were talking about what was going to happen in a basic situation of war reporting: A firefight, say, occurs at Point A, and cameramen and photographers rush to Points B through Z to cover it. Under the rules of embedment, this is not supposed to happen. Each cameramen and photographer, just as each reporter, is to be assigned to a specific unit, and is supposed to stay with that unit unless permitted to leave -- and, anyway, none of the embedded journalists is permitted a vehicle, so as to enable him or her to run off to Point A from B through Z. In embedment theory, the cameraman attached to the unit engaged in the firefight is supposed to get the picture, and everyone not attached to that unit is supposed to stay where he or she belongs and not get the picture.

Ryu was of the fairly firm view that this arrangement could not work. There was no way, she said, that people for whom the picture was the story (and for whom the picture of the war could be a career-maker) were going to sit still and miss the picture, no matter what the rules said.

Blumenfeld did not argue the point. "Okay, it is a problem," he said. "But we don't have a fleet of taxis. . . . We are trying to deal with what we can, as we can. This embedment process is the best we can do to work out a compromise. . . . It is something we will have to talk about as we go on."

In the first Gulf War, in which Blumenfeld served as a public affairs officer, the U.S. military, in collaboration with the major American media companies, built a system that was designed to sharply limit direct observational reporting to a relative few journalists, overwhelmingly drawn from the ranks of big media. The permitted few were to file "pool" reports and pictures that would be made available to all media through a military clearing process. Unsurprisingly, the arrangement turned out to please no one. The coverage was spotty and shallow, with the majority of American reporters covering the war from hotels and briefing rooms; one reporter's inevitably subjective view of an event that only he had covered was of little use to colleagues trying to craft an "objective" account several hundred miles away. Much was lost to history. And, of course, reporters who wanted to report the war for themselves simply went off on their own.

The experiment -- "the huge experiment," as Blumenfeld says -- this time represents an admirable attempt to do much better. And it would seem that it must be better: A system that allows eyewitness reporting across the spectrum of conflict, no matter how constrained, has to produce a picture of war, and of the military that goes to war, more true and complete than a system that seeks to deny eyewitness reporting.

But Blumenfeld's honest response -- "it is something we will have to talk about as we go on'' -- will, I think, turn out to be something of an understatement. There are problems of control and independence that are unresolved here, and these, as Ryu suggested, will come roaring up the first time they meet the first practical test.

The Department of Defense ground rules for embedding speak of the imperative "to tell the factual story, good or bad." For the sake of that great goal, I hope the Pentagon thinks more about loosening things up a bit. But also, I hope so for the sake of the military's media front line, public affairs officers like Blumenfeld. As any White House press secretary can tell them, there is no hell quite so annoying as the hell of an infantilized media pack.