Immanent laws of information

We are publishing one of the articles of the recently deceased outstanding Russian computer scientist, one of the creators of the legendary KAISSA, Mikhail Donskoy, found in his papers.

The force of gravity existed before Newton. All the people on earth felt it well, but could not formulate what it was. In the same way, there are objective laws of information that work regardless of our desire. Attempts to go against these laws end as painfully as the flight of Icarus.

I am not Newton and cannot formulate these laws, but I would like to express some thoughts on this matter. I apologize in advance to the scholars for using my own terminology, since different sources call the same things differently.

Information is a natural phenomenon, and it exists always and everywhere. Each person perceives information, generates new information and transmits information to other people. We can say that there are information nodes that receive information from the outside, process it, generating new information, store some part of it and, finally, transmit some part of the received or generated information to other nodes.

The receipt and transmission of information are not instantaneous, but permanent, so we need to talk about information channels between information nodes. An information channel is a channel for transmitting information between two specific information nodes.

Information nodes can be not only people, but also animals, computer programs, systems, organizations, devices, and so on. We are within the framework of a global information system in which information nodes constantly appear and disappear, and information channels between nodes are constantly formed and deleted. The nodes themselves, in turn, can be considered as independent information systems with their own subnodes and channels between them.

By the way, an important special case of transmitted information are control signals in control systems. In general, most information control systems are created with the goal of improving the quality of control, which in the above diagram should be understood as the quality of information in control signals.

Each node perceives information received through its input channels in accordance with its properties. For example, a person does not perceive ultrasound, but perceives color, and bats - exactly the opposite, perceive ultrasound, but see poorly.

When processing information inside a node, its properties also play a significant role. A poorly written computer program may not calculate the formula correctly. It is human nature to understand other people’s words to the extent of their depravity, and so on.

The storage of information received or generated by the node itself depends on the amount of available memory, storage and search methods and, thus, also depends on the properties of the node.

Finally, the transmission of information to other nodes is dictated by the available transmission channels, and, in the case of people or human groups, their interests.

From this enumeration it is clear that the information node distorts information out of necessity. Something is not perceived, something is forgotten, something is incorrectly assumed, and somewhere there is a deliberate distortion of information due to the interests of the node. There is no desire to send reliable information to some channels, and to some there is fundamentally no such possibility. Thus, we can talk about the quality of an information channel as the coefficient of information distortion in this channel.

Another important property of an information channel is the speed of information transmission. It significantly depends on the physical properties of the channel through which the channel is transmitted.

In addition, in the case when the receiving node is a person, the method of presenting information, roughly speaking, its understandability, plays a big role. It is known that graphs are perceived by humans better than a sequence of numbers. It is necessary to distinguish between the method of presenting information (how the same information is transmitted - either as a graph or as a sequence of numbers) and the transmission of different information about one entity (selected indicators).

Progress is constantly taking place along the path of organizing new information channels in order to speed up the transfer of information and improve its quality. However, it should be remembered that the most important nodes in information systems were and remain people who have their own limitations, shortcomings and interests. There is a well-known formula for information systems - “Garbage in, garbage out.” If people for some reason generate poor-quality information, then no means of transmitting and processing information will make it correct, and control signals generated on the basis of this information will be ineffective.

This, alas, does not mean that control signals generated on the basis of high-quality information will be effective, because some kind of distortion still occurs, including in the node generating the control signal. Translated into everyday language: if the boss is a fool, then honest subordinates will not improve the matter.

It is worth adding to this that some information channels, by definition, cannot be of high quality. Such channels include, for example, vertical channels, namely channels between a boss and a subordinate. This gives rise to problems in managing large teams.

There are many different sciences in different areas of human knowledge that consider the construction of information systems as the organization of a channel structure. In addition to the obvious problems of information technology - the topology of computer networks, the organization of communication systems and so on, there is a humanitarian branch - the theory of organizational management, political science and others.

Everywhere the task is the same. How to build an effective information system, having a given set of information nodes or a limited choice of nodes of certain types. Which nodes to choose, how to connect them with channels for effective management of this system. This problem is solved by taking into account the speed and quality of channels.

If you build a system for managing a team of people based on horizontal information channels, that is, having a small number of management levels, the top management level will have too many incoming channels and will not have time to work with them. And if they are built according to the vertical principle, that is, with a large number of levels, then on the way to the highest management level the information will be distorted beyond recognition. In this context, the “power vertical” is a guarantee of ineffective decisions.

It was very interesting for me to observe how the management of large IT firms in the USA in the 90s, with the growth of companies, was rapidly approaching the management schemes in the USSR. The apotheosis was the reform of Apple Computer carried out by D. Scully. This was painfully reminiscent of Evstigneev’s role in “Beware of the Car.”

An interesting theory by A. Levenchuk is that managers should first be taught programming as a way of organizing information systems, and then move on to methods of organizing human teams.

See also:

• Mikhail Blinkin. “He always did exactly what was interesting to him”
  
• Mikhail Donskoy. Life cycle of a programmer