What is "Quasi-Random"?
I sincerely hope that all of my fellow High Programmers are adept enough at programming to recognize what "Quasi-Random" means: the appearance of randomness, or sharing characteristics with randomness, without being truly random. Quasi-random sequences of numeric values are employed in many algorithms, from the simple selection of INFRARED Morning Morale Enhancement Music to immense combat simulations run by Armed Forces.
Why use Quasi-Random numbers?
True Randomness isn't as easy to come by as you might think when looking at R&D, the average price of ChapStik, or YELLOW clearance fashion trends - while many things seem Truly Random, they rarely are. The True Randomness used in Alpha Complex is carefully extracted from highly radioactive materials and the upper atmosphere of Outside, and is much more potent than Quasi-Randomness. In modern times, the use of True Randomness is generally reserved for Friend Computer's subsystems and certain models of 'bot brain - though there are exceptions. In Old Reckoning times, True Randomness was not much used, except as an execution method for beasts called "cats".
Where do Quasi-Random numbers come from?
There are two broad categories of Quasi-Randomness: algorithmic and observational. Algorithmic Quasi-Randomness uses a mathematical formula with an initial seed value to create an arbitrary sequence of numbers. Observational Quasi-Randomness attempts to better simulate True Randomness by observing some aspect of the world that's fairly random, such as the quantity of Algae Muffins consumed in the last 30 secondcycles, or the number of citizens visible to security cameras in a particular corridor. The top 10 sources of Quasi-Randomness (by queries per unit time) in Alpha Complex are:
- The Realtime INFRARED Injury Report Index;
- Fractal Hair Pattern Disassembly;
- The Lollman-Hertz Equation;
- Power line fluctuations in Sectors ZAP, ZOT, OHM, and WAT;
- Standardized Pipeline Turbulence Measurements;
- The (Blinks / Belches) derivation;
- Volumetric Scrubbot Cargo Modeling;
- Mad-MAN monitoring;
- Treason-O-Meter Mk 7.2
What problems exist in the system?
- The fundamental difficulty is that even the Observational sources aren't really random. They're run through filters to remove as much predictability as possible, but it is undeniable that some correlation remains between the activity in question and the generated numbers. In order to assist programmers with selecting an appropriate source of Quasi-Randomness, these correlations are well-documented, which leads to the second problem: tampering. Anyone who knows exactly how, say, the Standardized Pipeline Turbulence Measurements work could conceivably try to influence programs based on that measurement by interfering with pipeline turbulence - or, more feasibly, altering the sensors which measure it. Fractal Hair Pattern Disassembly contains a trivial vulnerability; if more than 5 sectors worth of citizens suddenly go bald, any number of programs will hang. Most methods have similar problems, of some variety or another. Algorithmic sources are not so vulnerable to tampering, but can be completely and easily reconstructed from their seed values. This makes them invaluable for some purposes, but unsuitable for others.
What in Friend Computer's name does this have to do with the Toothpaste Disaster?
Quasi-Random algorithms and processes are in use throughout Alpha Complex. There are three areas in which Quasi-Randomness (or True Randomness) is more than peripherally relevant to the Toothpaste_Disaster:
The 12th most popular source of Quasi-Randomness is the Amalgamated Baseline_Dental_Health Index. It is possible, if unlikely, that some of the events of the Toothpaste Disaster were performed in an effort to influence this Index - though such a task would be quite difficult, and probably require extensive work on the part of at least two High Programmers.
All four reactors in the Four_Reactor_Foam_Flood were generators of True Randomness used by Friend Computer. This is the case for only 2.17% of all arbitrary four-reactor proximate clusters in Alpha Complex.
Close examination of the logs for Treason-O-Meter Mk 7.2 reveals a recurring pattern. If decoded into standard alphanumerics, it reads, "The Wall_Will_Fall".
This last item is particularly disturbing, as it is strong evidence that treasonous parties have been using this particular algorithm (based off the timestamps of treasonous acts) as a communications medium, implying either total compromise of the generated data or an unprecedented level of coordination of treasonous elements. Needless to say, Treason-O-Meter ought to be swiftly investigated and shut down.
I dunno 'bout no wall fallin' or reactor randomness, but I can tell youse for a fact dat it's possible t'singificantly effect d'Realtime INFRARED Injury Report Index. Easy, even, if'n youse know d'right people. Which I'm not sayin' I do, mind. Interested parties what might be wantin' t'follow up on some'a dis-- purely f'academic purposes, understand-- should contact me through offline channels t'set somethin' up. Ask around OUS Sector; maybe some folks know me, maybe not.