Containment (12 page)

The message remained in the corner of his workspace for the next several hours, and Arik became increasingly curious. It wasn't uncommon to see diagnostic output for a few seconds or maybe even a few minutes while someone tried to track down a problem with the live system, but he'd never seen something like this remain visible for an entire day. He was thinking of contacting someone in the Code Pod when he got a video message from his father asking him if he had time to look into what he called the "anomalous string" that was appearing in the corner of everyone's workspace. Darien seemed to be in a hurry, and sent off the message without any additional information or details. Arik looked at the time and realized that Cadie would be home from work within the hour. He knew that they would have to discuss the baby very soon, but now that he had a new problem that needed solving, it wouldn't have to be tonight.

Arik wondered why the request to debug the problem had come from his father. Darien was a chemical and structural engineer. He headed up the Wet Pod and had designed several of the buildings in V1. Like all engineers, he knew computers well, but he didn't have any obvious stake in bugs in the shell program. He was good friends with Fai, however, which suggested to Arik that Fai had probably asked Darien for his son's help. Fai would have been too proud to ask Arik for help directly, and Arik imagined that the circuitous request through his father was still presented more as the Technology Department simply not having the time or resources to be distracted by such a trivial issue. But if the request did in fact originate from Fai, that meant the message was not simply diagnostic output, but probably a series of error codes that were unusual enough that nobody in the Code Pod had any idea what they meant.

Arik stood up in front of the polymeth wall and stretched while bringing up the source code for the shell program. He had been taking pain medication all day, and he needed to stand and move around the room in order to clear his head and stay focused.

Before he even had a chance to begin his debugging ritual, he recognized the first number in the error code, 2519658000000, as a date. Since computers weren't inherently able to distinguish one absolute date from another, they used relative dates expressed as some unit of time since a known epoch. V1CC inherited the ancient convention of expressing moments in time as the number of milliseconds since midnight on January 1, 1970. Since numbers like 2.5 quadrillion didn't come up very often in day-to-day computing tasks, when they did, it was usually safe to assume that they were machine-readable dates. And since the last six digits were all zeros, Arik could even tell that the number probably pointed either to exactly noon, or exactly midnight.

The date was most likely what programmers referred to as a "time stamp." Error codes almost always came with time stamps so whoever was debugging the problem could figure out exactly how long ago it happened. But when Arik did the math of subtracting the error code's time stamp from a time stamp representing the current time, he was surprised to find that the result was a negative number. The computer wasn't reporting a problem that occurred in the past; it was predicting an error 2.75 days in the future.

Although computer models were used to predict the probability of errors and failures all the time, as far as Arik knew, V1CC was not programmed to perform predictive diagnostics on itself. It was far more likely that the computer's clock had wandered prior to printing out the message, or was even wandering now. As powerful as computers were, left to their own devices, they were astonishingly lousy timekeepers. In order to keep their internal clocks accurate, they needed frequent calibration. Every 90 minutes, V1CC received a signal from a satellite that passed overhead which contained one of the most accurate clocks ever built. The clock used 12 lasers to monitor the optical light emitted by the electrons in a single atom of ytterbium. Counting the tiny pulses of light allowed the clock to break a second down into almost a quadrillion parts. By the time the Sun burned through most of its hydrogen gas and expanded to the point that all life in the solar system was destroyed, the ytterbium clock would have likely strayed less than one second. Of course, for V1CC to benefit from the accuracy of their micro-gravitational optical atomic clock, it would have to successfully receive the time calibration signal.

Arik instinctively checked his watch which consisted of two separate dials: a digital module which calibrated with V1CC, and an analog mechanical movement which used a steel spring, rotor, gear train, escapement, and about 200 additional parts to keep time to within a few seconds a day without relying on any external power source or time calibration signal whatsoever. Although mechanical watch movements were mostly favored by obsessive and anachronistic hobbyists, several of the computer scientists in V1 found them useful for keeping tabs on V1CC. There was no way a mechanical watch could detect a fraction of a second drift in V1CC's timekeeping, but it could detect a loss or gain of time adding up to a couple of minutes or more. When things like the life support system relied on the computer maintaining almost perfect time, and the computer relied on an atomic clock orbiting 12,000 kilometers above the surface of the planet, it seemed like a good idea to have some kind of an isolated analog backup.

But both times on Arik's watch agreed to within a few seconds, and a quick review of the logs showed that V1CC had only missed a handful of time calibrations in Arik's entire lifetime, the last one being over four and a half years ago. Whatever the time stamp meant, it was probably accurate.

Arik ran the shell program inside of another program which could trace the rendering of each pixel back to the exact line of code that initiated the drawing instruction. He drew a rectangular debug region around the message in the lower right-hand corner of his workspace, and restarted the shell. He found that the message was being rendered by a little over a hundred lines of code interspersed throughout the shell's source, nestled in among other similar lines of rendering code so seemingly randomly that it had to have been done intentionally. Each component of the message was calculated using a long and complex equation. Some of the variables in the equations were even random numbers, yet each formula was orchestrated in such a way as to somehow compensate, always yielding the exact same result.

Now that Arik was sure that the message was intentionally injected into the shell program, he believed it had to be an attempt to communicate with someone inside of V1 — very possibly him. He looked at the second and third numbers again, and now that he had a fresh perspective, he recognized them instantly. They were radio frequencies. 922.76 MHz was the frequency the Earth Radio Pod used to communicate with the satellites that relayed signals to and from Earth, and 40.002 MHz was the frequency that V1 used to communicate with the ERP. The ERP was isolated from V1 so that in the event of a catastrophic accident, it might still remain functional. It was a small structure only large enough for one or two people, and it was located a full kilometer south of V1 where it was well out of range of fires or shrapnel should the unthinkable occur. It had its own computer system, power supply, and miniature life support system based on tanks of compressed air. The only connection between V1 and the ERP was the 40.002 MHz radio link.

Two radio frequencies and a date three days in the future suggested to Arik that the message wasn't so much a message in and of itself as it was instructions on where and when to find the real message. The problem was that Arik wasn't able to listen in on either of those two frequencies. All communications to and from Earth were highly secured using encryption algorithms that Arik would be hard pressed to break anytime in the next decade, even with a multi-core electron computer. That, Arik believed, was what explained the word "DELTA." In the context of radio communication, "delta" was usually used in place of the letter "D," however an alternative interpretation — the variation of a variable or function, or the difference between two values — seemed to make much more sense. The difference between the two encrypted frequencies was 882.758 MHz — a frequency which, as far as Arik knew, wasn't being used for anything, and which he should be able to easily tune in to using the V1 frequency scanner.

By this time, Arik was almost positive that the message was intended for him. He was also fairly certain that it was either a trick being played on him by a friend of his in the Code Pod, or possibly a test arranged by Dr. Nguyen or Priyanka to make sure Arik was still up to the task of solving AP. He checked the source control system's logs to see who was responsible for the changes to the shell program, and was astounded to find that all of the revisions had been attributed to him.

This was almost certainly not a joke. Embedding "easter eggs" in code for fun and covering your tracks was one thing, but attributing changes to another user was much more difficult, and in the case of Arik's account, very nearly impossible for anyone except maybe Fai himself. Not only did Arik use the standard DNA identification protocol, but he was probably the only one in V1 who combined biometric identification with gesture identification. Gesture identification required that unique shapes or patterns be drawn in order to verify someone's identity. Even if someone had figured out how to spoof his biometric signature, his gesture ID was complex enough that it couldn't be guessed, and since he almost always used his BCI to draw it, it was unlikely that someone could have covertly recorded it, or deduced it from marks or prints left on a piece of polymeth. The likelihood that Arik's account had been compromised was extremely low.

It was far more likely that Arik's memory of hiding the easter egg had been destroyed either by the accident, or in the surgery afterwards, and that the message was an attempt to pass along information to himself in the future. The theory made perfect sense except for one thing: it implied that he had somehow been able to predict the accident.

S
ince Arik and Cam began their careers and got married, they very rarely saw each other. They both had so much to learn in their respective fields that neither of them had much time for anything outside of work. What little time and energy they had left at the end of the work day usually went into maintaining their marriages rather than their friendships. All four of them had managed to get together twice for dinner and a little four-handed chess, but both evenings ended prematurely: one night, Cam kept nodding off in his chair and drooling down the front of his shirt, and the other, Cadie fell into a deep sleep on a convertible futon between turns and had to be carried into the bedroom. She woke up the next morning asking who was ready for dessert.

So Arik and Cam decided that they would get together occasionally during the day. Although they wouldn't be able to spend much time away from work, at least they could spend that time talking rather than trying to stay awake. But lunch proved much trickier to coordinate than dinner. Between emergencies, midday meetings, having to eat with senior colleagues, and simply feeling too overwhelmed to get away, they found that their schedules refused to align. Although Cam was as overworked as anyone in V1, it was usually Arik who sent the terse last-minute cancellation message which Cam had learned to check for on the nearest piece of polymeth before leaving the Wrench Pod and boarding the maglev.

The fourth time Arik canceled, he sent Cam a long and detailed apology. As usual, he was sorry for the short notice. He wasn't taking their friendship for granted, and he knew that he needed to work on his priorities. But he was feeling tremendous pressure to make some sort of a breakthrough. Subha was constantly asking for updates. AP was turning out to be a much more difficult problem than he anticipated, and he was starting to wonder if it would end up being his life's work rather than just his first assignment. Again, he was sorry.

Cam later joked that they could have easily had lunch in the time it took Arik to write the apology. But Cam's immediate response was just a single line which Arik never forgot:

"Let's be the only two people in V1 who never have to say they're sorry."

It was Cam who then came up with the idea of spontaneous scheduling. Rather than the futile exercise of trying to anticipate a day on which they would both be free only to have to repeatedly cancel when something unexpected came up, the assumption would always be that they were both too busy. If one of them discovered that they were able to get away, they would send the other a message by 1145 hours. There was no need to respond if you couldn't make it, and no need to apologize or justify yourself. Just try again another time.

The system worked. They had each sent a couple of unacknowledged messages, and neither felt guilty. And then on a day when Subha didn't come in and there was the light and carefree atmosphere of a holiday around the Life Pod, Arik went into his office to send Cam a message only to find that Cam had already sent one to him.

"I'm free. You?"

For some reason, it was always assumed that if their schedules ever aligned, Cam would come to the Life Pod and they would eat there. But Cam had seen the Life Pod before, and had even gotten a rare personal tour of the dome. Nobody ever told Arik that he wasn't allowed to show people around, so he felt reasonably comfortable pleading ignorance if it turned out to be against department policy. Fortunately, nobody seemed to mind, though it probably didn't hurt that he conducted the tour at 2300 hours, long after all of Arik's colleagues (including Subha) had gone home. All the automatic lighting in V1 had faded in accordance with sunset in Aksai Chin (coordinating day and night between V1 and the GSA's headquarters meant more sleep for everyone), and Cam was expecting the Life Pod to be dark. But there were still hundreds of hours of daylight remaining in the Venusian solar day, so the mustard yellow sunshine filled the dome, penetrated the polymeth airlock, and lit up almost the entire Life Pod. Even the hallway was bright enough that they had to squint and shield their eyes as they walked toward the dome. With no windows in V1, it was almost impossible to imagine that the Sun did not rise and set with the rhythms of the human race.