Thoughts
current edition of the article on the Makeloft website
What is a thought? How do deaf people think from birth? What dreams do the blind see? And what is the world like for blind and deaf people? How does memory work? Why do we remember objects from a computer point of view more complex for perception (faces, pictures, melodies) than elementary (phone numbers and long sequences of characters)? Where does consciousness begin in the chain of living beings? What are emotions and feelings? Is artificial mind creation possible? Consciousness?
The article is exploratory in nature and contains simplified hypotheses, but we will try to at least a bit closer to understand these issues. Maybe someone will know their thoughts or find new ones ...
')
Associative-relational data model
This area has not yet been sufficiently studied, but some patterns are of interest and can be used in practice.
Let's start from afar, with solving everyday tasks: you need to create a more or less detailed database of the university, including a schedule of individual and group classes, audience numbers, lists of students and staff, while describing the structure of the university (educational groups, faculties, departments, positions and other attributes) ).
The classical relational approach involves the allocation of a large number of interrelated entities, for example, a student, group, subgroup, faculty, stream, department, employee, position, occupation, subject, type of occupation, time, place, audience, type of audience and many others. With the addition of each new entity, the database schema becomes more and more complex and tied to a specific educational institution, that is, applying it to another, say, where the hierarchy of educational subgroups or the schedule principle is slightly different, is not so easy. That is, the universality of the subject area is lost ...
Is it possible to create something more generalized? If you carefully analyze the subject area, you can still see that most of the entities are abstract in nature and only a very small specific, namely, specific: Person (individual person), Place (any place), Event (atomic time event) . All the rest is an attempt of the human mind to classify (organize, cluster) sets of specific objects into logical groups ...
Then the data model can be described approximately as follows:
PS If suddenly someone wishes to go deep, there are the source codes of a workable Dispatcher study program written a few years ago on the basis of the principles below.
Now let's look at what the features, disadvantages and advantages of such a scheme are.
For example, we want to describe students and university staff. To begin with, we add a “Status” entry to the PersonClusterType table, which reflects the status of a person. In our case, there may be two such statuses “Student” and “Employee” - we will add these two records to the PersonClusterValue table, and they will refer to the “Status” type. That is, we have already formed two clusters of people, "Students" and "Employees", and one person may well join both clusters, that is, study at the university and combine work in it.
Great, we figured it out. Now it is logical to assume that employees have posts. In the PersonClusterType table we add the entry “Position”, and in PersonClusterValue a number of entries with the names of the posts “Assistant”, “Technician”, “Laboratory Assistant”, “Teacher”, “Head of Department”, “Dean”, “Rector”, “Secretary” Watchman ”,“ Janitor ”,“ Watchman ”,“ Dressmaker ”and others who refer to the record of the type“ Position ”. In this case, one person can also simply combine several such posts.
We now turn to the students. As a rule, they are divided into groups, subgroups, streams and faculties. Create a “Study group” entry in PersonClusterType, and add a lot of groups to PersonClusterValue, for example, “750503”, “750504” and so on ... After this step, it’s very easy to get a list of people who study in a particular group. With subgroups the situation is more interesting - within the same group there can be subgroups on physical education, a foreign language, on laboratory studies.
In the PersonClusterType table, add the “Foreign Language Subgroup”, “Physical Education Subgroup” types and so on, and the instances of these subgroups themselves in the PersonClusterValue table, and then associate them with the corresponding main training group (many-to-many closed connection).
Similarly, for example, create the “Faculty”, “Department” and “Stream” types in the PersonClusterType table. And in the PersonClusterValue table we add the appropriate instances of the faculties - KSIS, FITU, IEF, FRE, FKP and so on, the departments - computers, POIT, Higher Mathematics, History, Philosophy, Psychology and others, and 75050, 85050 and many others in the flows.
After that, taking advantage of the many-to-many closed connection at the PersonClusterValue table, we associate the training groups with the relevant departments and streams. Similarly, you can associate and some employees in faculties and departments.
Using this approach, you can absolutely arbitrarily group and organize people, as well as create complex hierarchies. This applies not only to people, but also to temporary events (Event) and places (Place). Moreover, connections acquire an associative character, which makes it possible to apply not only a linear search by records, but also a real associative one. And this, in turn, opens up new opportunities for caching and processing large data arrays.
Moreover, if we look at it in more detail, then grouping can be done in part in various ways, and the same DB scheme can be applied not only to the university, but also, say, to a commercial company or school, which gives it unprecedented versatility.
And how would the scheme look like a subject domain in the classical relational representation?
Summarizing what has been said, the method consists in the fact that in separate database entities we only carry out concrete objects of the real world (for example, Person), and for related abstract concepts we create only an entry in the PersonClusterType table. In a conventional relational approach, we would have to create a separate table for each such abstract record. As a result, it turns out that we endow a certain set [cluster] of Person objects with a new property of the [PersonClusterType] type and assign a certain value [PersonClusterValue] to this property.
It is worth noting that, in terms of programming, the properties attached to objects in this way are of string type, but nothing prevents to improve the proposed scheme and use different types, for example, Boolean or numeric. However, with numbers there are limitations, which will be mentioned below.
Due to the closed connection of many to many, these attached properties (clusters) are capable of having a complex character, not atomic, that is, to represent a hierarchical structure.
Moreover, you may have noticed that the principle of filling such a database somewhat resembles the learning process ... And the program itself can simulate thinking algorithms, ask questions or even guide the user ... For example, when we add a new person to the Person table, based on the fact that almost all users have any status of “Student” and / or “Employee”, the program itself is able to suggest the user to assign a new person to some of these clusters (deductive reasoning from the general to the particular ). Then the application will offer to correlate the employee with the position, and the student with the study group.
On the other hand, it is possible that all educational groups are not related to faculties. We associate one or two educational groups with faculties, and the program suddenly “guesses” that other educational groups also need to be correlated with faculties and proposes to perform this action for them (inductive inference from private to general).
Moreover, even more complex types of conclusions are successfully formalized on this type of database. For example, objects X and Y are included in clusters A, B, C, D. At some point, the user creates a new cluster E and adds an object X to it. The program, revealing an analogy between X and Y, also suggests adding Y to a new cluster E.
Associative data model
If we go further into this approach, then those column properties that Person has, namely FirstName, LastName, Birthday and others, can also be rendered into clusters, that is, the table will eventually turn into “pure entities-images”, where an object other than its Id is no longer a column, and to restore its state, it is necessary to collect the values ​​of its properties for all associated clusters.
From the standpoint of OOP in such a situation, there is a complete semanic gap, therefore, it will hardly be convenient, at least very unusual, to program something on such a database. But still, if we present objects in the form of “property-value name” dictionaries, then something interesting can be done, namely, it becomes quite simple to implement algorithms of inference.
But if you go even deeper ... Erase the difference between the entities Person, Event, Place (create the type “Entity” in ClusterType, and add three entries in ClusterValue with this type “Person”, “Event”, “Place”) and leave only three table entities:
- clean images of instances of objects (Image)
- values ​​of properties of clusters of objects (ClusterValue)
- types of properties of clusters of objects (ClusterType)
It turns out that such a scheme can be described absolutely any kind of objects. From the Image table, you need to take a specific instance and see which clusters it belongs to, on the basis of which, to restore all its properties and their values. Interesting, isn't it? However, this is not the limit - everything can be minimized in general into two entities-tables Cluster and Value (in OOP, everything is an object, and in associative databases everything is a cluster). Of course, in this case it will not be easy to retrieve instances of cluster image objects from the table, but here an intriguing thing is hidden - it all becomes like a pattern of image associations in the brain ...
But back to the numbers. In real databases, mathematical data is often stored in the fields of table entities, for example, floating point numbers. If you use a purely associative approach, then in the ClusterValue table you need to add the whole set of numbers, and this is just a huge array of data in real databases. Obviously, it is extremely difficult for the effective operation of the entire table and the search by associations. But this fact, as it were, hints at us about why people often remember poorly sequences of numbers and symbols.
As for the practical applicability of the approach with modern OOP programming, it suffices to dwell on the associative-sequential level and not go to the fully associative one. But we will go further.
Perception model
We will try to present in a simplified form how some of the senses work.
Let's start with a view. As soon as light hits the retina, signals from receptors are formed and, based on them, the brain begins to select primitive images: lines, color areas, areas of light and shadow. Gradually, the order of images becomes more complicated and now, according to information from the previous stage, geometric forms begin to form, then the brain tries to fold more complex objects from a variety of such geometric figures - from a balloon to another person’s face, for example. If the object cannot be recognized immediately, then the brain begins to look for analogies using the properties of this object, so the child may well call the balloon a ball. Or, the process of external learning is initiated - a question is asked to an adult.
Children's drawings help to understand this scheme even more, where quite complex objects of the real world are depicted as primitives, but nevertheless are perceived as what is needed. This is the power of art - to convey images of a higher order (emotions, feelings) with the help of more elementary ones, but so that it reaches another person ...
As for hearing, the process is somewhat similar. It starts with the fact that the auditory receptors receive information about the frequencies contained in the sound signal, after which the information is collected for a short period of time and changes in it are analyzed, after which a kind of drawing is formed ... It is by him that we recognize the melody played on various instruments and recognize words uttered by completely different voices.
In fact, during wakefulness and partly when we sleep, the brain continuously performs such work of recognizing surrounding images and clustering in the background. As soon as something important comes into view, attention is transferred to this object and its analysis begins to occur in more detail.
The perception of primitive images from the sense organs weakly depends on the emotional state of the individual, although it may, for example, strengthen or weaken in certain situations, but the clustering of high-order images is very strongly subordinated to emotions.
Emotional model
As a rule, most of the recognized object images of the world have a neutral emotional tint, but in some circumstances and states of an individual, this color becomes bright and may even translate the individual to another emotional state.
The nature of the individual in the simplification can be represented in the form of a graph, where the vertices are all his possible atomic emotions and each is matched by the current level of arousal. All emotions are connected by edges, where each edge has its own conductivity coefficient (the probability of arousal transfer from one emotion to another). The values ​​on the edges are stable over long time intervals, and the values ​​at the vertices are dynamic, and easily influenced by images of the external world. Some emotional states are most probable and therefore stable (analogous to a potential well), and significant external or internal stimuli are necessary to move into others.
Mental model
Since people have the main channel of communication and cognition of speech (oral and written) in the brain, a habit of pronouncing thoughts is developed. Perhaps this is necessary as an exercise in order to be in good shape and quickly formulate phrases. More complex channels — visual and emotional — are usually not so frequently used, so imaginative thoughts are less common. However, they allow you to think associatively, bypassing a strict logical level and processing information faster, because speaking takes time and control. Perhaps someone knows the state of "flashed the thought," probably this is an example of pure figurative thinking, when the brain establishes a new relationship, but did not have time to arrange it in the form of a speech or visual design.
Obviously, deaf-and-dumb since childhood do without speaking, using only visual concepts, so their ways of thinking are different from those familiar to most people. Blind people from birth cannot use visual images, but their tactile and auditory channels are well developed. It is noteworthy that people who are blind in adulthood retain the ability to dream, however, as experiments on animals show, the ability to see is laid in early childhood, and if a newborn animal is glued with one eye and removed after a long time, the animal will remain blind. on this eye, and the sight on it will not recover, although the entire visual channel will not have any obvious damage. Just in the brain will not establish associative links responsible for the recognition of elementary images from this channel.
Since childhood, deaf and dumb since childhood are extremely limited in their ways of learning, communication and thinking, therefore, without due attention from other people, their mental development stops. However, as living examples show, with the right approach, deaf-blind people can successfully master even a higher education.
Consciousness
Where does consciousness begin? Most likely, a single body cell, a virus or a bacterium does not possess consciousness, and it is also unlikely that a separate part of the body is endowed with consciousness, even with nerve fibers. Intuitively, consciousness is associated with the brain as a whole. Moreover, an interesting moment, as a rule, in one copy ... Perhaps there are exceptions when several consciousnesses appear in one brain (split / reproduction of personality) or consciousness is lost altogether (vegetative state), but here you need to be careful, since these deviations can characterize one mind.
What is it all about? Result of nervous activity or "divine spark"? It is worth noting that reality exists for consciousness continuously from the moment of birth to death, and for reality the presence of consciousness is an interrupted process (birth, sleep without dreams, temporary loss of consciousness, death).
The materialistic deterministic scientific approach says that no supernatural entities can exist and everything is limited by the laws of the real world. , ( ), , , . …
, , , , , . , , , , ?
, , — () . « », , , , , . , , , , - . , , , ?
, , , , ( , ), , , , . , . . , () . , , , , !
, , , .
. ( ) . — , . , , .
, ? , , , … , ? , … , - , , — ! , , … . , — . , , , - - , . !
, , , , ( ), … , , , , ! , , - , . , - -, .
42
, , , , , , !
What is a thought? How do deaf people think from birth? What dreams do the blind see? And what is the world like for blind and deaf people? How does memory work? Why do we remember objects from a computer point of view more complex for perception (faces, pictures, melodies) than elementary (phone numbers and long sequences of characters)? Where does consciousness begin in the chain of living beings? What are emotions and feelings? Is artificial mind creation possible? Consciousness?
The article is exploratory in nature and contains simplified hypotheses, but we will try to at least a bit closer to understand these issues. Maybe someone will know their thoughts or find new ones ...
')
Associative-relational data model
This area has not yet been sufficiently studied, but some patterns are of interest and can be used in practice.
Let's start from afar, with solving everyday tasks: you need to create a more or less detailed database of the university, including a schedule of individual and group classes, audience numbers, lists of students and staff, while describing the structure of the university (educational groups, faculties, departments, positions and other attributes) ).
The classical relational approach involves the allocation of a large number of interrelated entities, for example, a student, group, subgroup, faculty, stream, department, employee, position, occupation, subject, type of occupation, time, place, audience, type of audience and many others. With the addition of each new entity, the database schema becomes more and more complex and tied to a specific educational institution, that is, applying it to another, say, where the hierarchy of educational subgroups or the schedule principle is slightly different, is not so easy. That is, the universality of the subject area is lost ...
Is it possible to create something more generalized? If you carefully analyze the subject area, you can still see that most of the entities are abstract in nature and only a very small specific, namely, specific: Person (individual person), Place (any place), Event (atomic time event) . All the rest is an attempt of the human mind to classify (organize, cluster) sets of specific objects into logical groups ...
Then the data model can be described approximately as follows:
PS If suddenly someone wishes to go deep, there are the source codes of a workable Dispatcher study program written a few years ago on the basis of the principles below.
Now let's look at what the features, disadvantages and advantages of such a scheme are.
For example, we want to describe students and university staff. To begin with, we add a “Status” entry to the PersonClusterType table, which reflects the status of a person. In our case, there may be two such statuses “Student” and “Employee” - we will add these two records to the PersonClusterValue table, and they will refer to the “Status” type. That is, we have already formed two clusters of people, "Students" and "Employees", and one person may well join both clusters, that is, study at the university and combine work in it.
Great, we figured it out. Now it is logical to assume that employees have posts. In the PersonClusterType table we add the entry “Position”, and in PersonClusterValue a number of entries with the names of the posts “Assistant”, “Technician”, “Laboratory Assistant”, “Teacher”, “Head of Department”, “Dean”, “Rector”, “Secretary” Watchman ”,“ Janitor ”,“ Watchman ”,“ Dressmaker ”and others who refer to the record of the type“ Position ”. In this case, one person can also simply combine several such posts.
We now turn to the students. As a rule, they are divided into groups, subgroups, streams and faculties. Create a “Study group” entry in PersonClusterType, and add a lot of groups to PersonClusterValue, for example, “750503”, “750504” and so on ... After this step, it’s very easy to get a list of people who study in a particular group. With subgroups the situation is more interesting - within the same group there can be subgroups on physical education, a foreign language, on laboratory studies.
In the PersonClusterType table, add the “Foreign Language Subgroup”, “Physical Education Subgroup” types and so on, and the instances of these subgroups themselves in the PersonClusterValue table, and then associate them with the corresponding main training group (many-to-many closed connection).
Similarly, for example, create the “Faculty”, “Department” and “Stream” types in the PersonClusterType table. And in the PersonClusterValue table we add the appropriate instances of the faculties - KSIS, FITU, IEF, FRE, FKP and so on, the departments - computers, POIT, Higher Mathematics, History, Philosophy, Psychology and others, and 75050, 85050 and many others in the flows.
After that, taking advantage of the many-to-many closed connection at the PersonClusterValue table, we associate the training groups with the relevant departments and streams. Similarly, you can associate and some employees in faculties and departments.
Using this approach, you can absolutely arbitrarily group and organize people, as well as create complex hierarchies. This applies not only to people, but also to temporary events (Event) and places (Place). Moreover, connections acquire an associative character, which makes it possible to apply not only a linear search by records, but also a real associative one. And this, in turn, opens up new opportunities for caching and processing large data arrays.
Moreover, if we look at it in more detail, then grouping can be done in part in various ways, and the same DB scheme can be applied not only to the university, but also, say, to a commercial company or school, which gives it unprecedented versatility.
And how would the scheme look like a subject domain in the classical relational representation?
Summarizing what has been said, the method consists in the fact that in separate database entities we only carry out concrete objects of the real world (for example, Person), and for related abstract concepts we create only an entry in the PersonClusterType table. In a conventional relational approach, we would have to create a separate table for each such abstract record. As a result, it turns out that we endow a certain set [cluster] of Person objects with a new property of the [PersonClusterType] type and assign a certain value [PersonClusterValue] to this property.
It is worth noting that, in terms of programming, the properties attached to objects in this way are of string type, but nothing prevents to improve the proposed scheme and use different types, for example, Boolean or numeric. However, with numbers there are limitations, which will be mentioned below.
Due to the closed connection of many to many, these attached properties (clusters) are capable of having a complex character, not atomic, that is, to represent a hierarchical structure.
Moreover, you may have noticed that the principle of filling such a database somewhat resembles the learning process ... And the program itself can simulate thinking algorithms, ask questions or even guide the user ... For example, when we add a new person to the Person table, based on the fact that almost all users have any status of “Student” and / or “Employee”, the program itself is able to suggest the user to assign a new person to some of these clusters (deductive reasoning from the general to the particular ). Then the application will offer to correlate the employee with the position, and the student with the study group.
On the other hand, it is possible that all educational groups are not related to faculties. We associate one or two educational groups with faculties, and the program suddenly “guesses” that other educational groups also need to be correlated with faculties and proposes to perform this action for them (inductive inference from private to general).
Moreover, even more complex types of conclusions are successfully formalized on this type of database. For example, objects X and Y are included in clusters A, B, C, D. At some point, the user creates a new cluster E and adds an object X to it. The program, revealing an analogy between X and Y, also suggests adding Y to a new cluster E.
Associative data model
If we go further into this approach, then those column properties that Person has, namely FirstName, LastName, Birthday and others, can also be rendered into clusters, that is, the table will eventually turn into “pure entities-images”, where an object other than its Id is no longer a column, and to restore its state, it is necessary to collect the values ​​of its properties for all associated clusters.
From the standpoint of OOP in such a situation, there is a complete semanic gap, therefore, it will hardly be convenient, at least very unusual, to program something on such a database. But still, if we present objects in the form of “property-value name” dictionaries, then something interesting can be done, namely, it becomes quite simple to implement algorithms of inference.
But if you go even deeper ... Erase the difference between the entities Person, Event, Place (create the type “Entity” in ClusterType, and add three entries in ClusterValue with this type “Person”, “Event”, “Place”) and leave only three table entities:
- clean images of instances of objects (Image)
- values ​​of properties of clusters of objects (ClusterValue)
- types of properties of clusters of objects (ClusterType)
It turns out that such a scheme can be described absolutely any kind of objects. From the Image table, you need to take a specific instance and see which clusters it belongs to, on the basis of which, to restore all its properties and their values. Interesting, isn't it? However, this is not the limit - everything can be minimized in general into two entities-tables Cluster and Value (in OOP, everything is an object, and in associative databases everything is a cluster). Of course, in this case it will not be easy to retrieve instances of cluster image objects from the table, but here an intriguing thing is hidden - it all becomes like a pattern of image associations in the brain ...
But back to the numbers. In real databases, mathematical data is often stored in the fields of table entities, for example, floating point numbers. If you use a purely associative approach, then in the ClusterValue table you need to add the whole set of numbers, and this is just a huge array of data in real databases. Obviously, it is extremely difficult for the effective operation of the entire table and the search by associations. But this fact, as it were, hints at us about why people often remember poorly sequences of numbers and symbols.
As for the practical applicability of the approach with modern OOP programming, it suffices to dwell on the associative-sequential level and not go to the fully associative one. But we will go further.
Perception model
We will try to present in a simplified form how some of the senses work.
Let's start with a view. As soon as light hits the retina, signals from receptors are formed and, based on them, the brain begins to select primitive images: lines, color areas, areas of light and shadow. Gradually, the order of images becomes more complicated and now, according to information from the previous stage, geometric forms begin to form, then the brain tries to fold more complex objects from a variety of such geometric figures - from a balloon to another person’s face, for example. If the object cannot be recognized immediately, then the brain begins to look for analogies using the properties of this object, so the child may well call the balloon a ball. Or, the process of external learning is initiated - a question is asked to an adult.
Children's drawings help to understand this scheme even more, where quite complex objects of the real world are depicted as primitives, but nevertheless are perceived as what is needed. This is the power of art - to convey images of a higher order (emotions, feelings) with the help of more elementary ones, but so that it reaches another person ...
As for hearing, the process is somewhat similar. It starts with the fact that the auditory receptors receive information about the frequencies contained in the sound signal, after which the information is collected for a short period of time and changes in it are analyzed, after which a kind of drawing is formed ... It is by him that we recognize the melody played on various instruments and recognize words uttered by completely different voices.
In fact, during wakefulness and partly when we sleep, the brain continuously performs such work of recognizing surrounding images and clustering in the background. As soon as something important comes into view, attention is transferred to this object and its analysis begins to occur in more detail.
The perception of primitive images from the sense organs weakly depends on the emotional state of the individual, although it may, for example, strengthen or weaken in certain situations, but the clustering of high-order images is very strongly subordinated to emotions.
Emotional model
As a rule, most of the recognized object images of the world have a neutral emotional tint, but in some circumstances and states of an individual, this color becomes bright and may even translate the individual to another emotional state.
The nature of the individual in the simplification can be represented in the form of a graph, where the vertices are all his possible atomic emotions and each is matched by the current level of arousal. All emotions are connected by edges, where each edge has its own conductivity coefficient (the probability of arousal transfer from one emotion to another). The values ​​on the edges are stable over long time intervals, and the values ​​at the vertices are dynamic, and easily influenced by images of the external world. Some emotional states are most probable and therefore stable (analogous to a potential well), and significant external or internal stimuli are necessary to move into others.
Mental model
Since people have the main channel of communication and cognition of speech (oral and written) in the brain, a habit of pronouncing thoughts is developed. Perhaps this is necessary as an exercise in order to be in good shape and quickly formulate phrases. More complex channels — visual and emotional — are usually not so frequently used, so imaginative thoughts are less common. However, they allow you to think associatively, bypassing a strict logical level and processing information faster, because speaking takes time and control. Perhaps someone knows the state of "flashed the thought," probably this is an example of pure figurative thinking, when the brain establishes a new relationship, but did not have time to arrange it in the form of a speech or visual design.
Obviously, deaf-and-dumb since childhood do without speaking, using only visual concepts, so their ways of thinking are different from those familiar to most people. Blind people from birth cannot use visual images, but their tactile and auditory channels are well developed. It is noteworthy that people who are blind in adulthood retain the ability to dream, however, as experiments on animals show, the ability to see is laid in early childhood, and if a newborn animal is glued with one eye and removed after a long time, the animal will remain blind. on this eye, and the sight on it will not recover, although the entire visual channel will not have any obvious damage. Just in the brain will not establish associative links responsible for the recognition of elementary images from this channel.
Since childhood, deaf and dumb since childhood are extremely limited in their ways of learning, communication and thinking, therefore, without due attention from other people, their mental development stops. However, as living examples show, with the right approach, deaf-blind people can successfully master even a higher education.
Consciousness
Where does consciousness begin? Most likely, a single body cell, a virus or a bacterium does not possess consciousness, and it is also unlikely that a separate part of the body is endowed with consciousness, even with nerve fibers. Intuitively, consciousness is associated with the brain as a whole. Moreover, an interesting moment, as a rule, in one copy ... Perhaps there are exceptions when several consciousnesses appear in one brain (split / reproduction of personality) or consciousness is lost altogether (vegetative state), but here you need to be careful, since these deviations can characterize one mind.
What is it all about? Result of nervous activity or "divine spark"? It is worth noting that reality exists for consciousness continuously from the moment of birth to death, and for reality the presence of consciousness is an interrupted process (birth, sleep without dreams, temporary loss of consciousness, death).
The materialistic deterministic scientific approach says that no supernatural entities can exist and everything is limited by the laws of the real world. , ( ), , , . …
, , , , , . , , , , ?
, , — () . « », , , , , . , , , , - . , , , ?
, , , , ( , ), , , , . , . . , () . , , , , !
, , , .
. ( ) . — , . , , .
, ? , , , … , ? , … , - , , — ! , , … . , — . , , , - - , . !
, , , , ( ), … , , , , ! , , - , . , - -, .
42
, , , , , , !
Source: https://habr.com/ru/post/250135/
All Articles