Strolling through informatics #1 – What is informatics

previous #0 ––– next #2

by Enrico Nardelli

(versione italiana qua)

Providing an exact definition of what a scientific discipline is can be a challenging task, partly because the very scholars in the field – when asked – produce different formulations. For informatics, the task is even more difficult, since this term, over the decades, has gradually come to encompass an enormous spectrum of concepts, from the most technological and operational ones (e.g., remote connection between two digital devices) to the most abstract and scientific ones (e.g., proving the mechanical solvability of a problem). This word is used, to give some examples, when referring both to technological devices (such as tablets or smartphones) on which various apps are run (i.e., the computer programs we use to search for information or to play games) and to the system software that enables the functioning of such devices, as well as to the formulas and theorems that prove the correctness of computational processes.

In this series of posts, I deal with informatics as a science, and we know that every science studies specific phenomena. For example, mathematics studies quantities, measurements and their relationships, physics studies non-living matter and its behavior in space and time, biology studies living organisms and their evolution.

What does the science called informatics study then? To guide the reader nonetheless, here is my definition: Informatics is the discipline that studies the automatic processing of representations.

Let us now analyze the terms of the definition in more detail, beginning by discussing what is meant by the adjective automatic. We are well acquainted with the concept of automation. Starting from the Industrial Revolution that began in the eighteenth century, it is quite clear to everyone that automating a production process means having machines do it. Products that were once made by hand, artisanally, are manufactured by machinery.

As a practical example, think of a woolen garment. Once upon a time, one had to shear sheep, spin wool, make cloth, then cut it and finally sew it, and all of this was done by people. Over the centuries, all these steps have been automated and are now performed by machines. Whenever we talk about a "machine" we mean a device that behaves mechanically, that is, when placed in the same starting conditions it always invariably performs the same operations, without possessing any awareness of what it is doing, thus without knowing the meaning of either what it is manipulating (or processing) or the manipulation (or processing) operations it is performing.

The adjective "automatic" in the definition is essential to characterize the fact that informatics deals only and exclusively with those situations in which processing occurs "mechanically," that is, deterministically and not subject to that consciousness and free will that instead characterize human beings. The technical term used in informatics to indicate a device that performs automatic processing of representations is "automaton," from the Greek word autòmaton, which is the root of "automation" and "automatic." A PC or tablet, smartphone or smartwatch are technologically sophisticated realizations of very complex automata. We will also use more informal terms like "agent" or "executor," always referring to an automaton.

By the term processing I appeal to common sense that refers to a manipulation of something, a transformation of something into something else.

In the automation example previously described, sheep's wool is transformed, that is, processed, through a series of different steps until a garment is obtained. In a more abstract context, that of numbers, the addition operation (3+5) transforms or processes the addends 3 and 5 into their sum (8). Indeed, informatics processing is also called "computation," since, in a historical progression, the first application scenarios were those related to numerical calculations.

The noun representations may raise some perplexity, since, generally, when talking about informatics, words like "data" or "information" are used. Deferring to a later post an in-depth examination of the meaning of these terms and their relationships, for now I clarify that by "representation" I mean the way – arbitrary but shared within some community of people – by which we make concrete in the physical world a datum, that is, a generic element present to our consciousness, regardless of whether it refers to objective facts or subjective acquisitions.

Returning to the example of numbers, we observe that they have no physical reality in themselves. The concept of "five things" can be physically represented by any set of five objects, but is abstract in itself. The representation we use is "5," but for the ancient Romans it was "V." Other populations use still different representations. We learn in our early school years, not without some effort, to understand the meaning of number representations and to do calculations with them. In this case, we too, as human beings, perform a processing of representations, but it is not entirely automatic, because we are people and not mechanisms.

In subsequent posts in this series, we will first explore the concept of representation more deeply, because it is precisely these representations that are processed by the executor. Subsequently, we will understand how this agent (the automaton) can be made and in what language (more precisely, programming language) we must provide it with the instructions we deem necessary to conduct the processing. We will then proceed with another fundamental concept, relating to the specification of what the processing procedure (technically, the algorithm) we want to realize should be. We will then consider the case of processing performed not by a single executor but by a set of them, a scenario called distributed computation. Finally, we will discuss a fundamental characteristic of informatics, abstraction, which, while not being an exclusive aspect of this discipline, since every science constructs its own models of phenomena through a process of abstraction from details, allows for the realization of machines capable of simulating other machines (virtualization). As a final concept we will address what, from a general cultural point of view, is perhaps the most important legacy of informatics, namely the duality between representations and programs.

The second part of the series will be dedicated to addressing some of the issues resulting from the increasing spread of systems and devices based on informatics.

[[The posts in this series are based on the Author's book (in Italian) La rivoluzione informatica: conoscenza, consapevolezza e potere nella società digitale, (= The Informatics Revolution: Knowledge, Awareness and Power in the Digital Society) to which readers are referred for further reading]].

--
The original version (in italian) has been published by "Osservatorio sullo Stato digitale" (= Observatory on Digital State) of IRPA - Istituto di Ricerche sulla Pubblica Amministrazione (= Research Institute on Public Administration) on 25 September 2024.