BEFORE WORLD WAR II, the great object of espionage was war plans: to find out whether or how a prospective enemy intended to attack. During the war, the focus necessarily shifted to codes and orders of battle, matters that could affect the course of the struggle. After the war, the urgency of the question of nuclear war or blackmail produced a new priority: the secrets related to the making and deploying of nuclear weapons. The backwardness of Soviet science gave the Kremlin an extra incentive to pursue this mission, even as the openness of American society gave it an extra opportunity. In fact, the Soviet effort to steal nuclear secrets began when the two countries were wartime allies. The Walker case is the latest evident sign of it.

The nagging question remains: how to keep the secrets. There can be no single set of answers. The beginning of one set, however, is to recognize whom we are currently dealing with: not with highly educated, politicized elite figures of the sort familiar from British fiction and reality, and not with professional master spies of the "Rudolph Abel" mold, but with, essentially, ordinary servicemen and national security workers. High-tech defense creates a requirement for large numbers of them to write the programs, keep the logs, change the codes. There may be no better way to ensure their loyalty than the gritty and, yes, often intrusive attention to individual performance and vulnerability that is the classic routine -- though it cannot be performed routinely -- of counter-intelligence.

But whether this vital work can be conducted with even minimal effectiveness in the existing bureaucratic circumstances is uncertain. The one useful result of the Walker case may turn out to be to propel review of a hopelessly overgrown and encumbered security system, one that seems to distinguish poorly between necessary and less necessary secrets and among the different techniques needed to guard different kinds of secrets. The overall problem has to be broken down into parts. High tech, for instance, at once manufactures and dissolves secrets: the same interchange essential to develop and apply them exposes them to being lost. Surely, however, the joints at which leaks are likeliest can be better sealed. Reducing and rationalizing the kinds of secrets to be kept and the numbers of people with access to them should not be beyond the wit of humankind.

Meanwhile, the law must be vigorously enforced against those caught in its coils. If personal gain is a motive for breaking security, personal loss should be confirmed as a price.