Dynamic Knowledge Representation as a Formalization Conveyor for Manmade Systems With Useful Impulse

Introduction

In a wide range of man-made systems (technical, social, economic and probably others) there is a peculiar class, in which indivisible or desirably indivisible pieces of useful material, energy or information are being transmitted from specified sender to specified recipient, and different circumstances are present that may affect this transmission in a good or bad way. Henceforth, the transmission of a singular piece from sender to recipient (with all possible adventures on the way) will be referred to as impulse (with intuitive rather than specifically physical connotation). Depending on the type of system, this impulse may be: delivery of any material stuff; money transfer from one account to another; sale and purchase of a material good; processing of a data packet; one utterance in communication between two people, and the like. By the way, if to refer to physics, impulse is a key notion of transactional interpretation in quantum mechanics (Cramer, 1986).

In the computer age, an eminent example of such systems is database management systems (DBMS). In them, the impulse carrying useful information is termed transaction, or, precisely speaking, simple (single) transaction – an indivisible sequence of data processing operations (Vassiliou, 1980). Gray (1981, p. 1) defined it as “a transformation of state which has the properties of atomicity (all or nothing), durability (effects survive failures) and consistency (a correct transformation).” In 1983, Reuter and Harder added one more property, Isolation (how transaction integrity is visible to other users and systems), and put forth an acronym ACID as shorthand for these four key properties (Reuter & Harder, 1983). ACID requirements ensure the security of data. To meet them, the transaction should be accomplished at once, not in parts, and transfer the database from one correct state to another, or, if for some reason some actions in the transaction are impossible, the database must return to the initial state (transaction rollback).

Modern banking is based on financial DBMS, in which transactions mean money transfers. For them, the issues of data (i.e., money) security are especially urgent. Financial institutions are one of those who suffer the biggest losses from misuse of computer-based tools. Analysis of crimes shows that although the overall number of thefts and robberies generally decreases worldwide, the number of wholesale money thefts increases (HSE, 1996-2017). By this report, more than one million of bank transactions, by which cash is withdrawn and illegally transferred onto offshore accounts, take place every day, total daily loss amounting up to 37 million dollars (Harris, 2009). This emphasizes the urgency of transaction safety issues, which, in turn, depend on the transaction design. The latter is largely governed by the formalisms used to encode the DBMS architecture (i.e., the transaction calculus). These formalisms, desirably, should be applicable to the entire class of systems with useful impulse and thus may serve many more purposes than they were originally developed for.

However, to refocus these formalisms on other fields, it is insufficient just to state that the systems they are now scheduled for are of the same class as before. A precise relation between relevant components of the systems is needed that would enable formalization first at the qualitative, and then, supposedly, at the quantitative level. To inquire whether such relation exists, both systems should be represented in one formal language, adequate enough to capture both relevant similarity and relevant distinction between them.