3. Recovery
ERTS REFERRED TO THE PROCESS AS “DATA RECOVERY,” or sometimes as “data salvage.” The terms evoked images of deep-sea operations, and they were oddly appropriate.
To recover or salvage data meant that coherent meaning was pulled to the surface from the depths of massive electronic information storage. And, like salvage from the sea, it was a slow and delicate process, where a single false step meant the irretrievable loss of the very elements one was trying to bring up. ERTS had whole salvage crews skilled in the art of data recovery. One crew immediately went to work on the audio recovery, another on the visual recovery.
But Karen Ross was already engaged in a visual recovery.
The procedures she followed were highly sophisticated, and only possible at ERTS.
Earth Resources Technology was a relatively new company, formed in 1975 in response to the explosive growth of information on the Earth and its resources. The amount of material handled by ERTS was staggering: just the Landsat imagery alone amounted to more than five hundred thousand pictures, and sixteen new images were acquired every hour, around the clock. With the addition of conventional and draped aerial photography, infrared photography, and artificial aperture side-looking radar, the total information available to ERTS exceeded two million images, with new input on the order of thirty images an hour. All this information had to be catalogued, stored, and made available for instantaneous retrieval. ERTS was like a library which acquired seven hundred new books a day. It was not surprising that the librarians worked at fever pitch around the clock.
Visitors to ERTS never seemed to realize that even with computers, such data-handling capacity would have been impossible ten years earlier. Nor did visitors understand the basic nature of the ERTS information—they assumed that the pictures on the screens were photographic, although they were not.
Photography was a nineteenth-century chemical system for recording information using light-sensitive silver salts. ERTS utilized a twentieth-century electronic system for recording information, analogous to chemical photographs, but very different. Instead of cameras, ERTS used multi-spectral scanners; instead of film, they used CCTs—computer compatible tapes. In fact, ERTS did not bother with “pictures” as they were ordinarily understood from old-fashioned photographic technology. ERTS bought “data scans” which they converted to “data displays,” as the need arose.
Since the ERTS images were just electrical signals recorded on magnetic tape, a great variety of electrical image manipulation was possible. ERTS had 837 computer programs to alter imagery: to enhance it, to eliminate unwanted elements, to bring out details. Ross used fourteen programs on the Congo videotape—particularly on the static-filled section in which the hand and face appeared, just before the antenna was smashed.
First she earned out what was called a “wash cycle,” getting rid of the static. She identified the static lines as occurring at specific scan positions, and having a specific gray-scale value. She instructed the computer to cancel those lines.
The resulting image showed blank spaces where the static was removed. So she did “fill-in-the-blanks”—instructing the computer to introject imagery, according to what was around the blank spaces. In this operation the computer made a logical guess about what was missing.
She now had a static-free image, but it was muddy and indistinct, lacking definition. So she did a “high-priced spread”—intensifying the image by spreading the gray-scale values. But for some reason she also got a phase distortion that she had to cancel, and that released spiking glitches previously suppressed, and to get rid of the glitches she had to run three other programs.
Technical details preoccupied her for an hour, until suddenly the image “popped,” coming up bright and clean. She caught her breath as she saw it. The screen showed a dark, brooding face with heavy brows, watchful eyes, a flattened nose, prognathous lips.
Frozen on the video screen was the face of a male gorilla.
Travis walked toward her from across the room, shaking his head. “We finished the audio recovery on that hissing noise. The computer confirms it as human breathing, with at least four separate origins. But it’s damned strange. According to the analysis, the sound is coming from inhalation, not exhalation, the way people usually make sounds.”
“The computer is wrong,” Ross said. “It’s not human.” She pointed to the screen, and the face of the gorilla.
Travis showed no surprise. “Artifact,” he said.
“It’s no artifact.”
“You did fill-in-the-blanks, and you got an artifact. The tag team’s been screwing around with the software at lunch again.” The tag team—the young software programmers— had a tendency to convert data to play highly sophisticated versions of pinball games. Their games sometimes got sub-routed into other programs.
Ross herself had complained about it. “But this image is real,” she insisted, pointing to the screen.
“Look,” Travis said, “last week Harry did fill-in-the-blanks on the Karakorum Mountains and he got back a lunar landing game. You’re supposed to land next to the McDonald’s stand, all very amusing.” He walked off. “You’d better meet the others in my office. We’re setting advance times to get back in.”
“I’m leading the next team.”
Travis shook his head. “Out of the question.”
“But what about this?” she said, pointing to the screen.
“I’m not buying that image,” Travis said. “Gorillas don’t behave that way. It’s got to be an artifact.” He glanced at his watch. “Right now, the only question I have is how fast we can put a team back in the Congo.”
4. Return Expedition
TRAVIS HAD NEVER HAD ANY DOUBTS IN HIS MIND
about going back in; from the first time he saw the videotapes from the Congo, the only question was how best to do it. He called in all the section heads: Accounts, Diplo, Remote, Geo, Logistics, Legal. They were all yawning and rubbing their eyes. Travis began by saying, “I want us back in the Congo in ninety-six hours.”
Then he leaned back in his chair and let them tell him why it couldn’t be done. There were plenty of reasons.
“We can’t assemble the air cargo units for shipment in less than a hundred and sixty hours,” Cameron, the logistics man said.
“We can postpone the Himalaya team, and use their units,” Travis said.
“But that’s a mountain expedition.”
“You can modify the units in nine hours,” Travis said.
“But we can’t get equipment to fly it out,” Lewis, the transport master, said.
“Korean Airlines has a 747 cargo jet available at SFX. They tell me it can be down here in nine hours.”
“They have a plane just sitting there?” Lewis said, incredulous.
“I believe,” Travis said, “that they had a last-minute cancellation from another customer.”
Irwin, the accountant, groaned. “What’d that cost?”
“We can’t get visas from the Zaire Embassy in Washington in time,” Martin, the diplomatic man, said. “And there is serious doubt they’d issue them to us at all. As you know, the first set of Congo visas were based on our mineral exploration rights with the Zaire government, and our MERs are non-exclusive. We were granted permission to go in, and so were the Japanese, the Germans, and the Dutch, who’ve formed a mining consortium. The first ore-body strike takes the contract. If Zaire suspects that our expedition is in trouble, they’ll just cancel us out and let the Euro-Japanese consortium try their luck. There are thirty Japanese trade officials in Kinshasa right now, spending yen like water.”
“I think that’s right,” Travis said. “If it became known that our expedition is in trouble.”
“It’ll become known the minute we apply for visas.”
“We won’t apply for them. As far as anybody knows,” Travis said, “we still have an expedition in Virunga. If we put a second small team into the field fast enough, nobody will ever know that it wasn’t the original team.”
“But what about the specific personnel visas to cross the borders, the manifests—”
“Details,” Travis said. “That’s what liquor is for,” referring to bribes, which were often liquor. In many parts of the world, expedition teams went in with crates of liquor and
boxes of those perennial favorites, transistor radios and Polaroid cameras.
“Details? How’re you going to cross the border?”
“We’ll need a good man for that. Maybe Munro.”
“Munro? That’s playing rough. The Zaire government hates Munro.”
“He’s resourceful, and he knows the area.”
Martin, the diplomatic expert, cleared his throat and said, “I’m not sure I should be here for this discussion. It looks to me as if you are proposing to enter a sovereign state with an illegal party led by a former Congo mercenary soldier
“Not at all,” Travis said. “I’m obliged to put a support party into the field to assist my people already there. Happens all the time. I have no reason to think anybody is in trouble; just a routine support party. I haven’t got time to go through official channels. I may not be showing the best judgment in whom I hire, but it’s nothing more serious than that.”
By 11:45 P.M. on the night of June 13, the main sequencing of the next ERTS expedition had been worked out and confirmed by the computer. A fully loaded 747 could leave Houston at 8 P.M. the following evening, June 14; the plane could be in Africa on June 15 to pick up Munro “or someone like him”; and the full team could be in place in the Congo on June 17.
In ninety-six hours.
From the main data room, Karen Ross could look through the glass walls into Travis’s office and see the arguments taking place. In her logical way, she concluded that Travis had ”Q’d” himself, meaning that he had drawn false conclusions from insufficient data, and had said Q.E.D. too soon. Ross felt there was no point in going back into the Congo until they knew what they were up against. She remained at her console, checking the image she had recovered.
Ross bought this image—but how could she make Travis buy it?
In the highly sophisticated data-processing world of ERTS, there was a constant danger that extracted information would begin to “float”—that the images would cut loose from reality, like a ship cut loose from its moorings. This was true particularly when the database was put through multiple manipulations—when you were rotating 106 pixels in computer-generated hyperspace.
So ERTS evolved other ways to check the validity of images they got back from the computer. Ross ran two check programs against the gorilla image. The first was called APNF, for Animation Predicted Next Frame.
It was possible to treat videotape as if it were movie film, a succession of stills. She showed the computer several “stills” in succession, and then asked it to create the Predicted Next Frame. This PNF was then checked against the actual next frame.
She ran eight PNFs in a row, and they worked. If there was an error in the data handling, it was at least a consistent error.
Encouraged, she next ran a “fast and dirty threespace.” Here the flat video image was assumed to have certain three-dimensional characteristics, based on gray-scale patterns. In essence, the computer decided that the shadow of a nose, or a mountain range, meant that the nose or mountain range protruded above the surrounding surface. Succeeding images could be checked against these assumptions. As the gorilla moved, the computer verified that the flat image was, indeed, three-dimensional and coherent.
This proved beyond a doubt that the image was real.
She went to see Travis.
“Let’s say I buy this image,” Travis said, frowning. “I still don’t see why you should take the next expedition in.”
Ross said, “What did the other team find?”
“The other team?” Travis asked innocently.
“You gave that tape to another salvage team to confirm my recovery,” Ross said.
Travis glanced at his watch. “They haven’t pulled any-
thing out yet.” And he added, “We all know you’re fast with the database.”
Ross smiled. “That’s why you need me to take the expedition in,” she said. “I know the database, because I generated the database. And if you intend to send another team in right away, before this gorilla thing is solved, the only hope you have is for the team leader to be fast onsite with the data. This time, you need a console hotdogger in the field. Or the next expedition will end up like the last one. Because you still don’t know what happened to the last expedition.”
Travis sat behind his desk, and stared at her for a long time. She recognized his hesitation as a sign that he was weakening.
“And I want to go outside,” Ross said.
“To an outside expert?”
“Yes. Somebody on our grant list.”
“Risky,” Travis said. “I hate to involve outside people at this point. You know the consortium is breathing down our necks. You up the leak averages.”
‘‘It’s important,’’ Ross insisted.
Travis sighed. “Okay, if you think it’s important.” He sighed again. “Just don’t delay your’team.”
Ross was already packing up her hard copy.
Alone, Travis frowned, turning over his decision in his mind. Even if they ran the next Congo expedition slambam, in and out in less than fifteen days, their fixed costs would still exceed three hundred thousand dollars. The Board was going to scream—sending an untried, twenty-four-year-old kid, a girl, into the field with this kind of responsibility. Especially on a project as important as this one, where the stakes were enormous, and where they had already fallen behind on every timeline and cost projection. And Ross was so cold, she was likely to prove a poor field leader, alienating the others in the team.
Yet Travis had a hunch about the Ross Glacier. His management philosophy, tempered in his rain-dancing days, was always to give the project to whoever had the most to gain from success—or the most to lose from failure.
He turned to face his console, mounted beside his desk. “Travis,” he said, and the screen glowed.
“Psychograph file,” he said.
The screen showed call prompts.
“Ross, Karen,” Travis said.
The screen flashed THINKING A MOMENT. That was the programmed response which meant that information was being extracted. He waited.
Then the psychograph summary printed out across the screen. Every E RI S employee underwent three days of intensive psychological testing to determine not only skills but potential biases. The assessment of Ross would, he felt, be reassuring to the Board.
HIGHLY INTELLIGENT / LOGICAL / FLEXIBLE / RESOURCEFUL / DATA INTUITIVE / THOUGHT PROCESSES SUITED TO RAPIDLY CHANGING REAL-TIME
CONTEXTS / DRIVEN TO SUCCEED AT DEFINED GOALS / CAPABLE SUSTAINED MENTAL EFFORT /
It looked like the perfect description of the next Congo team leader. He scanned down the screen, looking for the negatives. These were less reassuring.