disappeared into numbers. Groteschele was explaining that of course there was a remote mathematical possibility of machine error. He had calculated the error. In any year the odds were 50 to 1 against accidental war. This, Black remembered, had already been made public in the Hershon Report. But the report was several years old and the people at Ohio State who put it together made it dear that they did not have access to confidential information. The situation was much worse than they had reported because everything had gotten more complicated.
“Putting it another way,” said Groteschele, “and with the present rate of alerts and the present computerized equipment, the odds are such that one accidental war might occur in fifty years.”
“Does not the increasing intricacy of the electronic systems and the greater speed of missiles make that figure worse each year?” Black asked. He had in mind the public warning, several years previously, by Admiral L D. Coates, the Chief of Naval Research, which admitted what all insiders knew: electronic gear was becoming so complex that it was outstripping the ability of men to control it; complexity of new generations of machines was increasing the danger of accidents faster than safeguards could be devised. The statement had never been countered but simply ignored.
Groteschele paused, smiled tolerantly at Black. The SecArmy leaned toward one of his aides and asked a question. The aide looked at Black, smiled as he talked quickly in SecArmy’s ear. Black knew what he was saying: General Black is the professional heretic.
“In theory, yes,” Groteschele said. “But we completely pretest every component and every system is checked by another. The chance of war by mechanical failure is next to zero.”
This was not true, but Black did not choose to contest it. He forced his anger down and deliberately looked at the Big Board. It was like a vast moving mosaic, decorative rather than functional. Blips appeared, grew bright, traveled short distances, then vanished. The Big Board was not taken seriously until the light over it indicated that someone had decided to come to some degree of alert.
In how many places in the world, Black mused, were there just such strategy boards run by just such computer systems, picking up, identifying, and discarding just such radar signals? Our big operation at Omaha, of course. Probably at least one other standby “Omaha” at some other part of the country. Then there was the President’s bomb shelter: probably another computerized strategy board layout there. Maybe there was more than one Presidential bomb shelter. Maybe one at Camp David, or the summer White House, or who knows where else.
Then, in addition, he knew there was always, every minute of every day, a converted KC-135 aloft: a miniatore, emergency “Omaha,” in case everything else should blow. Probably another one on some super aircraft carrier somewhere. Still another on one of the nuclear submarines. How many others? Several in England, certainly, France, perhaps, and West Germany. Russia? Yes, surely there would be almost as many as there are in the United States.
Black knew that four KC-135s were reserved exclusively for Presidential use as a flying command post in case of emergency. Since 1962 they had been scattered about the country so that the Preskknt was never far from one. Whenever the President flew overseas one of these planes was quietly and unobtrusively included in his escort.
Surely each one of them in every country was similar. And the men in them, too. Today, throughout the world in each one of the Big Board rooms, staffed by busy, competent, dedicated men, probably the same signals were being received, analyzed, and projected. The big brass everywhere was watching similar strategy boards, studying the same blips, thinking out, or talking Out, the same strategic puzzles Groteschele was now discussing.
Black plugged back into the discussion. Groteschele was now classifying various types of possible machine errors. Accidental war caused by some machine failure. Miscalculation by the computers, misinterpretation by the staff of human interpreters (the “overriders” as the computer boys call them). And then the big one, electronic failure.
Big, Black thought, because no one knows anything about it. We just know that in any system so complex and so dependent upon intricate electronic equipment, the possibility of electronic failure or error must always be borne in mind.
“But the Positive Control Fail-Safe system is the ultimate protection against mechanical failure,” Groteschele was saying, his voice heavily persuasive. This was the ultimate safety factor in the whole system. This was where it all rested. So, we were all reassured. And indeed, Groteschele was very reassuring today, except that his nervous chuckle kept getting in the way. Black understood why. Groteschele knew more than he was saying.
It just was not that simple. Everyone knew it who had anything to do with the black boxes of the Positive Control system. Their components were 100 per cent double-checked on regular rotation schedules. Every possible condition to which the equipment might be subjected in operation was simulated. It was simulated in actual duplication systems arid it was also simulated on computers with meticulously devised mathematical formulas expressing every possible way the equipment might fail. On computers the bombers were “flown” and the “go” signals were “given.”
Variable atmospheric conditions could be predicted, operational deterioration of the equipment could be estimated, vibration characteristics of the Vindicator bombers could be factored in. Stress variables could be translated into mathematical formulas, and with these formulas the computers could test out the black boxes.
But the whole system had one big flaw in it. Nobody could ever be certain that the black boxes would actually work properly in a showdown. The reason was simple. There had never been a showdown, and there could never be a sure test showdown.
A showdown meant war. The whole Positive Control system really depended on equipment that could never really be tested until the time came for its first use, and because of this nobody could ever really know in advance whether or not it would work right. The Fail-Safe machines could be truly tested only once: the single time they were used.
There was ample evidence from the experience of the Electra planes and the now obsolete DC-6s that a serious flaw in an elaborate machine could survive every experimental situation-and then in real practice come completely unstuck.
This was material for the grim inside humor which went the rounds of SAC gossip. The DC-6 had been a beautiful drawing-board plane, except that the first ones to go into service caught fire in flight. Then it was discovered that the one thing they hadn’t calculated was what flight wind currents would do to fuel overflow spillage. The fuel was’ deflected by an invisible band of air to a point directly behind the engines where the air- intake vents sucked in the gas spillage, converting the plane’s storage compartment into a quite unplanned fire chamber.
Great corporations can also be injured when their computerized positive control systems break down. Black remembered the consternation a few years back when Fortune had demonstrated this point about General Dynamics. The Convair 990 was a 200 million-dollar demonstration of the fallibility of computerized simulation. Convair designed a drawing-board airplane that checked out to be the fastest commercial jet in the computer “flight tests,” so they decided to save money, skip the costly prototype stage, and go directly into production. Only the 990 didn’t perform as designed. Nobody knew why, and the enigmatic computers that had been so reassuring could not be charged with malfeasance.
General Black also knew that Groteschele was sliding past another important factor. Each machine had to be adjusted and installed by men. And men, regardless of their training, suffered from fatigue and boredom. Many was the time that General Black had seen a tired and irritated mechanic turn a screwdriver a half turn too far, fail to make one last check, ignore a negative reading on a testing instrument. On a plane, such errors would mean only that an expensive piece of machinery and a few men would be lost. On a Fail-Safe black box-and the men who adjusted and installed them had not the remotest notion of what they were- the slightest accident could trigger the final disaster.
Black glanced around the table. Stark was watching the strategy board. New blips had just appeared and Stark was toying nervously with his pencil as he followed their progress. Black looked at Wilcox. He sensed that Wilcox was mentally rejecting the possibility of accidental war. It depressed Black, and he felt that he should do more, but he knew he could not.
Groteschele’s patronizing voice reached back into Black’s consciousness. Now the voice was talking about what if there should be an accident. An “interesting” problem, it was saying.
“Suppose the Russians caused an accident,” suggested Grotesdiele. “Suppose it were a true accident. Suppose it was a 50-megaton missile aimed for New York or Washington, what could be done? How could we really
