After only 45 minutes Valkyrie had reached a depth of 300 meters, where Martin was now watching the crew. This stop had been planned so they had enough time to check the course of the hole and the condition of the machinery. Martin’s help was not needed for this. The check wasn’t actually necessary, either, as the software would have issued an alert in case of any deviation. However, they were not going to blindly trust the programming.
Martin had interlaced his fingers and was twiddling his thumbs. One minute, two minutes, three minutes passed. He wondered, did my simulations take such a consciously triggered full stop into account? Probably not. He had particularly concentrated on catastrophes caused by the outside world. Suddenly, Martin shivered. Why did I not ask for today’s mission plan? He shook his head. I probably should not worry at all. Nevertheless, he would be even more careful next time, by a whole order of magnitude.
“Valkyrie to Flight Director, when is it going to continue?”
Francesca had asked this in such a way that an unspoken ‘finally’ could not be ignored.
“FD to Valkyrie, just one moment.”
This was the voice of Stone, who seemed to have personally taken over the role of Flight Director. It was very unusual that during a NASA program an external person served as FD, even though he was the inventor of this technology.
“FD to Valkyrie, radar showing an obstacle at two o’clock, X minus 20.”
“Confirmed. Is this a problem?” Francesca replied.
Was Stone getting overly cautious? Martin wondered about this, and pulled up a window with the radar on his display. Valkyrie should be able to get around this obstacle quite easily, should this even be necessary. They would see once they got closer. There would still be enough time for an evasive maneuver.
“FD to Valkyrie, no, just a cross-check.”
Something in Stone’s voice made Martin nervous. In the cockpit video stream, he could see Francesca had unbuckled her safety belt. Had she noticed something?
For a second, the image on Martin’s monitor wobbled. Valkyrie seemed to try to leap forward, but was held in place by something. Alarm messages were scrolling on the right edge of the screen.
“Valkyrie to FD, what the hell is going on? That really shook us up!”
Francesca seemed to be on the verge of screaming into the microphone.
“FD to Valkyrie, my apologies. We tried to restart the vehicle.”
“Without telling us? What is really going on there?”
“Sorry, we did not want to alarm you. Please be patient for a while!”
Martin shook his head. This certainly isn’t standard NASA procedure. Once this test is over, there will be repercussions. He already heard his colleagues bitching about the private partners. However, NASA could no longer maintain its research program without external investment.
“FD to Valkyrie, please report status.”
A useless request, as the monitors clearly showed all systems working within normal parameters.
“This is Valkyrie, status of all stations optimal.”
“FD to Valkyrie. There is a problem. Valkyrie refuses to start up.”
“It refuses?”
“It’s a safety protocol,” Stone said. “Some fool prohibited starting it if the radar detects an obstacle less than 30 meters away.”
This was a reasonable limitation—during the phase of diving into the ice Valkyrie could not be steered by software. The vehicle simply followed the laws of physics and that did not allow for evading an obstacle. The programmer had not considered one thing—there was a difference between a cold start at the surface and a restart in the middle of the ice, where Valkyrie was already in a vertical position. Martin was glad he was not the one who had programmed the system. Why hadn’t I simulated this condition, though? Because it did not seem threatening enough to me, Martin answered himself.
“FD to Valkyrie. No reason for concern. I am going to find a good programmer. I already have an idea.”
June 30, 2045, Antarctica
“So there’s enough oxygen for three hours. I am gradually starting to be worried.”
Francesca’s voice sounded calm and controlled. She’s either a great actress or really hard-nosed, Martin perceived.
“Yes, I am working on it.” He suppressed a don’t you worry that was on the tip of his tongue. The fact she had contacted Martin via the open network rather than through the direct connection with the Flight Director confirmed that she was worried—and she knew she had every reason for it—and he knew this, too. Despite this, the problem was actually quite trivial. Because the radar saw an obstacle in the direction of travel, the drill jets would not activate. This behavior made sense for a launch at the surface, but at a depth of 300 meters it might be deadly. Otherwise, the technology was completely functional. The laser supplied the vehicle with energy, so Francesca and Devendra were neither sweating, nor would they die of hunger or thirst. They could communicate with the entire world.
However, they would suffocate in about three hours. The expedition had been planned to last for only a few hours, but as of now it had gone on for almost two days. During this time, Martin had not slept, unlike the crew members, who had been told to nap as long as possible to save on air. He imagined how he would have reacted, stuck deep inside the ice. Would I have been able to sleep a wink? Up here, he at least had the feeling of being able to achieve something. He tried to ignore the responsibility resting in his hands, in his fingers that tried again and again to hammer new lines of code into the keyboard.
After all, he had not programmed the Valkyrie software. When Mission Chief Stone—who knew very little about programming—had first described the problem, Martin had still been optimistic. He had thought, How complicated could it be to outwit a safety mechanism that was obviously overreacting, like the immune system in the case of an allergy? The true degree of difficulty only emerged gradually. It had to do with the fact that he did not know the software well enough. While he had tested it in simulations, he did not understand the concept behind it. Being too familiar with the code was potentially bad for testing, because you ran the risk of ignoring bugs that ought to be prevented by the software architecture.
First, he had needed to grasp the concept, and Stone had provided him with all the material he needed. Martin could have logged into the NASA supercomputer, or even used it all by himself if this would have helped somehow. However, this was not a problem that could be solved by sheer computing power. Several times Martin had believed he found the decisive routines—and he had been wrong each time. The programming team did good work, he concluded. In the final analysis, the control software was supposed to be functioning ‘space-safe’ at the highest level of safety, without human intervention. Out there, the crew would be working light hours and millions of kilometers away from any human help.