“It’s on the menu. Just select it.”
“Got it. Big display, huh?”
“It’s meant to be used while you’re wearing flight gloves. You can change the size if you want.”
“Got it. So... this is a virtual version of my personal computer within the SAF tactical computer, right?”
“It’s an independently operating SAF personal computer. Not virtual. You’re accessing your office terminal by an individual domain that’s assigned to each SAF PC. That private domain acts as a single personal computer, so I suppose it’s a virtual computer if you look at it that way. You can load any sort of application software you have on it, but you can’t actually connect to the real PC through it. It was set up that way for security, so you can’t call it a totally virtual computer. Didn’t you know that?”
“I remember it being explained to me, but seeing it laid out in a hierarchical display like this is the first time I can really see how it works,” Captain Foss said as she looked at the EW display. “That means that all the data on Dr. Lecter is going through the tactical computer. I never realized that at all.”
“Fundamentally, it can’t be tampered with. I suppose anybody could if they got permission from General Cooley, but it’s probably pretty hard to get clearance under false pretenses. That’s the official line at least, but from a hardware perspective, the personal computers are just a segment of the tactical computer. The STC could probably do whatever it wanted to them if it decided to.”
“I’m sure. Like access it freely from Yukikaze the way we’re doing now.”
“The SAF super linker makes that possible. The SSL lets you operate the tactical computer from Yukikaze by means of a transmission protocol that’s unique to the SAF. You can’t do it unless you’re on a mission. We’re only able to do this because we’re acting under General Cooley’s orders.”
“But that means that any squadron member could do this. My patients’ privacy would mean nothing then,” Foss said. “I suppose encrypting my stored files would be a waste of time.”
“No, it wouldn’t. It’d take time to decrypt them, which greatly increases the likelihood of you noticing if someone’s been reading them on the sly. Anyway, let’s get to work.”
Captain Foss’s computer Dr. Lecter had profacting applications installed. Rei had learned of the various psychoanalytical and diagnostic software applications Foss used from his numerous examinations, but as to what specific tools she was using now, he hadn’t a clue. Rei monitored her work from the display in front.
Captain Foss started her profacting software.
“This is T-FACPro II, the most powerful profacting tool currently available,” she explained to Rei. “Using the numerical data it outputs from the effects of real-world behavior caused by mental load components, I can set up a certain situation and then simulate what sort of actions the target personality will take and the mental state they’ll have. It’s actually a standard tool in profacting. Ever heard of it?”
“No,” replied Rei.
“The psychoanalysis engine this tool runs on is top-notch, but it can’t really demonstrate its true effectiveness here in the FAF.”
“Why not?”
“T-FACPro II was designed to be linked up with MAcBB, a huge active database used exclusively for T- FACPro. MAcBB contains the prediction data from countless researchers who use T-FACPro II as well as the actual behaviors of the target personalities. When the gap between the predicted behavior and the actual behavior is too great, the prediction technique is judged to be flawed. Then, T-FACPro II’s analysis engine looks for the cause and attempts to select a prediction technique more in line with observed behavior. It’ll be hypothetical, but that hypothetical method would send feedback from the other versions of T-FACPro II running out there through the MAcBB database. If similar examples show that the method works, then it’s no longer considered hypothetical. By those means, we accumulate usable methods and samples through MAcBB. In other words, the greater the number of researchers accessing it, the smarter T-FACPro II’s analysis engine gets. However, I can’t access MAcBB from the FAF. Cut off from that, T-FACPro II can’t reach its full potential.”
“But it’s not completely useless, right?”
“Right. With my own know-how, we have a pretty good chance of getting a result. My T-FACPro II is better than the FAF’s own analysis tools and probably on par with the ones used exclusively by the SAF.” As she said this, Captain Foss began entering numeric data that seemed to be about Lieutenant Katsuragi.
“You said it compares predicted behavior with actual behavior?” Rei asked, being careful not to interrupt her.
“Right. Predictions made by the software are technically called predicts, while the actual behavior are actuals —”
“How does T-FACPro II know what the effects of real-world behavior from the target are? Do you input that?”
“Basically, yes. It isn’t always possible for it to judge if a prediction was accurate or not. If we take you for example, if I were to operationalize some specific behavior like ‘How will Captain Fukai feel about Yukikaze from now on,’ and profacted you, T-FACPro II would output an answer like ‘He will fear her and take steps to eliminate his fear.’ Then I’d keep you under observation to see if that’s correct, and after I judge if it is, I enter the result. After that, T-FACPro II will judge the method it used as being a good one.”
“But nothing you just described depends on that tool, does it?”
“That’s because you’re recognizing that the results of your profact are correct. That would be useful in an actual profacting where you’re making more detailed and specific predictions. The predicted result isn’t always the actual one, so it outputs a probability for how likely a prediction will be to come true.”
“But, in the end, it’s still just a prediction,” Rei said. “Basically, the program’s just a tool used by a researcher who’s accumulated experiences for the database. No matter how precisely T-FACPro II makes its predictions, there’s no guarantee that they’ll be correct.”
“Yes, of course. And having said that, there are times when it has such a high degree of accuracy that a user may become convinced that its predictions are absolutely correct. It’s easy for experts to fall into that trap, to say nothing of amateurs. Well, shall we give it a try? What would you like to know about Lieutenant Katsuragi? Try giving me a concrete scenario. For instance, something like ‘What will he say when he sees Yukikaze?’ ”
“Is that even possible to predict?”
“I still can’t profact him with a high degree of accuracy. All I did was enter his PAC code.”
“What’s a PAC code?”
“His standardized personal analytic classification code. Don’t you know about that? The FAF authorities used it to determine that assigning you to the SAF 5th Squadron would be a good fit for you. Naturally, Lieutenant Katsuragi had a code assigned to him as well. That’s what I was inputting just now.”
“Is everyone in the FAF numerically classified like that?”
“Not just in the FAF. You may not have known it, but you’ve probably had a PAC code following you around since childhood. Well, how they use the code varies from nation to nation. Some just use them to classify criminals, but the PAC code itself is a worldwide standard. It’s not unique to the FAF. Starting with blood type and an introversion/extroversion scale, it classifies you according to various psychological tendency components and arranges them as numeric data. T-FACPro II can read PAC codes directly, but in a full-scale usage of the software, it utilizes something called a PAX code, which is just an extended version of the PAC code, in order to achieve a greater level of predictive accuracy. It adds even more detailed psychological tendency components onto the PAC code, and coding those extensions is basically my job in all this. It leads to an even more detailed personality assessment. While you have to use T-FACPro II when you’re profacting, anyone can operate the software itself.”
“I see.”
“Since a mistake in analysis would lead to a flawed profacting, I’m sure you can understand why it requires an expert’s experience and advanced knowledge.”
“Have you made a PAX code for me too?”
“Yes, but your values aren’t fixed. You’re gradually changing.”
“I’d expect that. Humans are essentially analog. I doubt digital numeric data could express an individual personality,” Rei said.
“No,” Captain Foss replied. “It’s actually quite rare to have to revise past PAX code data to achieve an accurate profacting, as in your case. I’ve never seen anyone demonstrate personality changes as dramatic as yours. I originally thought I’d made a mistake when I first created your PAX code, but T-FACPro II determined that it was
