Information science and technology as applications of the physics of signalling

TL;DR

This paper proposes a physics-based theoretical model for information science and technology, extending signaling theory by framing facts and tests as physical phenomena, explaining system unpredictability and suggesting improved methods for software development.

Contribution

It introduces a novel physics-inspired model for information processes, linking tests, facts, and phenomena, and offers new methods applicable across the software life-cycle.

Findings

The model explains non-uniqueness in real-time system responses.

Restrictions on inference concurrency are justified by the model.

Improved software methods are summarized based on the model.

Abstract

Adopting the scientific method a theoretical model is proposed as foundation for information science and technology, extending the existing theory of signaling: a fact f becomes known in a physical system only following the success of a test f, tests performed primarily by human sensors and applied to (physical) phenomena within which further tests may be performed. Tests are phenomena and classify phenomena. A phenomenon occupies both time and space, facts and inferences having physical counterparts which are phenomena of specified classes. Identifiers such as f are conventional, assigned by humans; a fact (f', f'') reports the success of a test of generic class f', the outcome f'' of the reported application classifying the successful test in more detail. Facts then exist only within structures of a form dictated by constraints on the structural design of tests. The model explains why responses of real time systems are not uniquely predictable and why restrictions, on concurrency in performing inferences within them, are needed. Improved methods, based on the model and applicable throughout the software life-cycle, are summarised in the paper. No report of similar work has been found in the literature.