ARIA Systems Breakdown

This document provides critical information for those who are expected to work with an A-Type for a prolonged period of time or who are required to perform maintenance and systems management tasks in the event that the system encounters problems. The systems breakdown segments the systems within the A-Type itself where possible, as such every system present within the A-Type is designed to be completely interconnected with every other system while remaining entirely modular allowing for easy removal and assessment of each individual system.

Within the A-Type the Datacore provides a superior retrieval time compared to the neural-stack despite its poor association system¹⁾. Its function is to act as a database for the A-Type, storing data such as encyclopedias, applications, star-charts, and other declarative “fact” based memory. The design of the Datacore makes it an external storage device making it a non-critical component of the A-Type, this design choice has limited impact on the functionality however, as the A-Type so long as it has access to its Xtal, is connected to the Datacore with minimal latency.

This system functions as a translator in real-time for the Quantum, Binary, Quad, and Neural computing of the A-Type. The Translator is an immense feat of computer engineering being an absolutely critical part of the operation of the A-Type. Without the Translator, the A-Type is only able to communicate via the Xtal in debugging mode to avoid potential damage to the unit.

The Buffer is a component of the CNS that functions as a highly optimized pathway of neural systems providing storage for variables and considerations accessible by both the active and inactive psychology kernel component. In its functions the buffer operates between 300 and 1000 times the speed of the rest of the neural network. Due to the high performance needs of this part of the A-Type's systems, it is among the most expensive to produce and replace.

Another component of the CNS, the A-Type's Activity Monitor logs every step of the processing performed by the A-Type and is crucial for debugging purposes, allowing a technician to track and locate the source of an error. While this component is not necessary for standard operation and causes significant electrical drain during high-volume quantum operations, it is invaluable for recovery and maintenance operations and is not recommended to be disabled outside of the most dire of circumstances.

The Neural Stack composes the main body of the CNS, acting as the memory and neural processing store for the A-Type, blending both operations together into a seamless system.

The system includes sizable dedicated processors for both quantum and binary computation built into the core of the A-Type system. The quantum processor is optimized to grant the A-Type with an immense capacity for problem solving and logical deduction, while the binary processor provides an impressive calculation capacity.

A Signature assigned to the A-Type assures that the neural pathways have not been replicated from a stolen backup, prohibiting the direct cloning of A-Types without the use of highly specialized proprietary equipment capable of reading and then duplicating the signature hidden within the A-Type's CNS. An A-Type without a signature will experience an immense loss in performance and eventually shut down. The signature is a completely self-sustaining aspect of the A-Type in normal operating circumstances and requires no maintenance.

An optical brainstem, the Dataplexer is a highly adaptable piece of hardware capable of taking on up to 256 8-way parallel optical communication ports with zero parallel clock-shifting, or 1024 independent serial communication ports. The Dataplexer functions in tandem with the CNS to dramatically increase the reaction time of the A-Type.

Sensorial memory is a term that refers to a chunk of memory tied closely to the Translator that acts as a buffer for the A-Type's conscious experience as well as a completely accurate and objective log of events within the last 48 hours of operation. This memory may be transferred into long-term memory to be parsed and saved for future revision. This memory function allows the A-Type to provide an unbiased mathematical opinion and accurate judgment and estimation despite being at its core a neural system.

The Short-Term memory of the A-Type is memory which is vulnerable to being lost in the event of a major system error. This is due to need for the Short-Term memory to constantly be in a state of transition, changing within the A-Type to meet its needs. Primarily Short-Term memory works to search for patterns via the Recognizer and attempts to identify trends which allow for the A-Type to understand and anticipate events neurally, bolstered by the resources of a vast database and the computational power of a starship grade computer behind it. With these resources, the A-Type is capable of quickly identifying problems long before they fall out of hand. To aid with predicting events outside of normal parameters the Short-Term memory features a function which sporadically injects data randomly from unrelated sources and events to encourage truly dynamic assumptions and conclusions and avoid falling into a regular pattern.

Long-Term memory for the A-Type is an immensely complex system and is stored in a vast dynamically tagged database which is constantly being amended and added to, while also creating a complete back-log of every change ever made to the system. Long-Term memory stores declarative memory²⁾, opinions³⁾, and procedural memory⁴⁾ which are loaded as kernel modules and extensions.

The A-Type's Microkernel functions essentially as a twin kernel with one twin performing the typical actions of a kernel, while the other operates in tandem behind the scenes to pick up the pieces in the event of a malfunction. This allows the system to continue running with little to no loss in performance over long periods of time, barring the loss of any operation that could not be performed.

The Recognizer is a system which is tied directly into the quantum processor of the A-Type with the function of constantly searching for and logging unusual patterns and attempting to decipher their meaning. The Recognizer is not a consciously recognized or alterable process and is key to the function of the A-Type.

Within the A-Type's systems Extensions are effectively learned skills acquired through observation which are constantly being refined and amended. Due to their modular structure, Extensions can be quickly traded out and exchanged for new Extensions, evading the process of teaching the A-Type how to perform a complex operation.

The foundation of any Extension, a Module is a self-taught Extension and is applicable for skills ranging from riding a bike to neurally controlling an array of weapon pods. Like Extensions, Modules are modular and can be exchanged quickly. Due to the natural process of amendment, refinement, and optimization of the A-Type an older Module tends to be more effective than a younger Module. Unfortunately, Modules require a calibration time to settle into the A-Type's Response Selection System before the A-Type will have proper insight on how to most effectively employ a new Module.

A permanent component of the Kernel, it is a system used for recognizing and managing memory.

Tied directly into the problem solving and pattern recognition system, the sensorial system works closely with the neural components within the A-Type responsible for creating new neural pathways.

Of the A-Type's components, this is the most complex. The Psychology is effectively what makes the A-Type tick and gives it personality. This system can be calibrated to create a desired behaviour, however A-Types often insist on locking down the calibration function. For this purpose, a Psychology override exists in some A-Type models. Using the Psychology override is typically considered a violation of trust by the A-Type and can result in further complications.

This consists of operations performed on-top of other components.

The A-Type is designed to be capable of emulating a wide variety of computational systems, past and present, with hundreds of Lorath specifications loaded natively and with the storage capacity to support many more. The only true limit to an A-Type's emulation capabilities is access to a code base which the A-Type can interpret, or hardware limitations for systems which require computational power exceeding that available to the A-Type.

Software that is directly compiled upon the A-Type is few and far between, however the facility to create such software exists. One example is the creation of a subroutine for vanity behaviours which encourages the A-Type to wear clothing. Through software applications, other mannerisms can be trained.

The Modulus is a method for the A-Type to “explore” a device electronically which it is unfamiliar with and map out every function possible based on complex estimates to create a control driver or Module for the new system. This exploration is usually performed through external cables, these cables allowing connection of external components when it suits the A-Type's needs.

With time the Modulus learns about it's user's preferences and re-calibrates the device to best suit its user. This is considered as a separate system as the A-Type also re-calibrates its own body and software functions similarly. This function is limited, however, by the limitations of its organic body, and hardware.

The A-Type's software can be taught to “hook” into events and act upon them⁵⁾.

Software can be used to execute processor commands such as complex math; for example calculating FTL parameters, or similar actions. Commands are tied into Modules and Extensions.

Software can execute physical actions based on complex dependencies. Actions can be as simple as “scratch yourself” to complex precision based actions which describe the exact balance and angle of every joint in the body. For example; with proper modification, sniping an object the size of a tennis ball from several kilometres away. Actions are tied into Modules and Extensions.

Original article by Osaka current version transcribed by Eistheid