The concept of systems and designs. Their place in the design of information systems
After reading the comments on the previous article Classification of structures: examples and misconceptions about the classification of structures, I realized how much different we have about the term structures. When I wrote an article, it seemed to me that this term is interpreted quite simply. But, after reading the comments, I realized that I should talk about him separately.
The explanatory dictionary Ephraim defines two different concepts, which are denoted by one term construction:
Let's try to translate them into a formal language.
')
Since composition is a set, the first concept is translated as follows: a construction is a set of objects interconnected by connections. At the same time, judging by the definition, the objects must be man-made and inanimate. That is, it is impossible to imagine the Earth in the form of a construction, if you do not assume that it was made by aliens. DNA cannot be represented as a construct, unless this DNA is created by someone. That is, in the definition of the structure it is necessary to add that the objects are man-made. For example, a set of objects: {fuselage, wings, tail} consists of man-made objects, and, therefore, may be called a structure. The design called the plane. I note that in this context, an airplane is not an object, but a multitude of objects (fuselage, wings, tail}. You can call it a multitude of aircraft (k).
How many objects can be in a design? There is no answer to this question in the definition. But we can assume that their final number is greater than one, because the definition speaks of connections. In total, it turned out: the man-made set of objects created by man, the objects are combined by connections, the set is finite, the number of elements is more than one.
In this case there is no obligatory condition for the construction to have a name, or explicitly to indicate the object whose construction is being considered. You can simulate and nameless design.
The second concept of the term “construction” means the following: a construction is an object that can be represented as a set of objects. For example, since an aircraft as an object can be represented as a set of objects consisting of the fuselage, wings and tail, it is also called a structure. In this text, to distinguish the designation of an aircraft as an object from the designation of an aircraft as a set, we can write: an aircraft (o) can be represented as a set of objects — an aircraft (k).
Any object can be divided into parts. We do not know indivisible objects. That is, any object can be called a construction? Not. Because not every man-made object can be divided into man-made parts. For example, casting (pig), being a man-made object, can not be divided into man-made parts. Therefore, the disc can not be called a construction.
Analyzing the term "construction", we found one important feature of the language: the object and its construction are called by the same name. That is, the plane (o) and the plane (k) in everyday life is called by one name: the plane. It is clear that the object and the set of objects are different concepts. In the dictionary of Ephraim, these concepts are different, but in everyday life there is one name, and therefore, people often confuse them and cannot separate these two concepts, denoted by one term. The same problem was in the process approach, in which the concepts of function, functional structure, script, and so on. called one term - process. Because of this, it seemed to many analysts that the function and the script are one and the same.
The confusion, which arises from the fact that two concepts are called in one word, manifests itself in the answer to the following question: what is this or that object? Answers can be divided into two types:
The first type of definition gives the definition of concepts (o), and the second type - definitions of (k).
Another example: a train is a train of railway cars connected between itself and driven by a locomotive or motor car. In this context, the definition of a train (k) is given. It can be said that a train is a long vehicle for transporting passengers or cargo by rail. This is the definition of the train (o). Interestingly, in the dictionaries you can find definitions of both those and other concepts.
In everyday life, we do not notice the difference between such definitions. For example, a group of analysts is shown a production line layout. Each can see completely different pictures. One will see an object called “production line”, the other - a structure having the same name. Since the object and its construction are completely different concepts, they will see completely different things. Until they agree on a single look at this layout, they will talk about different objects. Well, if the context makes them converge on one point of view. However, this does not always happen. The stage at which it turns out the subject of discussion is usually skipped. Because of this, errors in understanding occur. The same problem arises when we want to build an ontological model. For example, if we want to find out the complexity of the aircraft design using the attribute: “the number of aircraft structural elements”, then we need to find an object in the model to which this attribute is attributed. It cannot be attributed to the plane (o), because it is possible to divide the plane into parts in many ways. Therefore, this attribute must be assigned to a set of objects, but not to an object.
Let's see how system engineering copes with this terminological paradox. System engineering gives the definition of a system like this:
We see that this definition is close to the definition of a construction in its first sense. That is, a system is a set of objects with properties:
The thesis about the man-made system is deleted from the list of properties. The system can be called not only a hand-made product, but also what exists in nature besides our participation. For example, biological systems. Therefore, the concept of a system is wider than the concept of construction. It can be said that constructions are a specialized subclass of the class of systems consisting only of man-made systems.
If for structures the relationship between an object and its structure was called so: “object construction”, then another term is used to designate the relationship between an object and its system: “object structure”. For example, the structure of a person connects a person (o) with a person (s). By the way, I wonder why there is no term “object system” by analogy with the term “object construction”?
Is it possible to call a system an object, and not a multitude of objects? That is, is it possible to apply the term system to an object in the same way as the term construct can be applied to an object? Most likely - you can. For example, it is said that the system has emergence. Formally, this thesis is translated as follows: the properties of an object, the structure of which is represented in the form of the system under study, are different from the properties of elements of this system. Since in this context an object is called a system, an object can also be called a system.
Since any object can be divided into parts, any object can be called a system. This distinguishes the term system from the term construction, because not any object can be called a construction.
It seems to me that in order to eliminate collisions that can arise for an engineer reading books on systems engineering, it is worth making the second definition of the term system into the dictionaries, by analogy with the second meaning of the term construction:
It would be nice if the texts on system engineering clearly indicate what kind of concept in this context refers to an object or a set of objects.
Is it possible to extend to the systems the thesis that any object can have different structures depending on the observer? Yes you can. We are well acquainted with two different paradigms of human structure, which give rise to different structures: the internal structure and external structure of a person.
In system engineering, there is also a requirement that imposes restrictions on the set of possible objects. It's about emergence. An object whose structure is represented as a system must have properties that are different from the properties of the elements of the system. This raises two questions:
Returning to the definition of a structure, one can ask the question: should the structure (o) have the properties of emergence? Suppose that should. Then the set of constructions (k) is a subset of the set of systems (c). If not, then the set of constructions (k) intersects with the set of systems (c).
Now let us try to generalize the concept of construction (k) and system (c) to a wider class of objects and sets. In my article I wanted to do just that. Apparently, without the current entry it was not clear. I introduced the concept of a generalized construction (k), which differs from the generally accepted concept of construction as follows:
In the last article I removed the word generalized from the name of the concept to be more concise.
The result is a hierarchy of classes: A generic structure is the widest set, of which systems and structures are a subset.
It took me to introduce the generalized structure to bring all the structures that we create to describe various structures, as well as to describe the limitations that arise when simplifying these structures, to a single view.
For example, most often, the modeling of structures is made using the “part-whole” links. In this case, information about the design (the average mass of structural elements, for example) we transfer to the model of the object, the design of which we model. The limitations of this modeling method are that we cannot create several different constructions of a single object, whether constructions in different paradigms, or constructions in the same paradigm, but differing in versions.
Once I was tasked with simulating different versions of the design of a single spacecraft. Versions existed simultaneously in time and modeled different versions of design solutions. In addition, the versions themselves changed over time, because design solutions evolved over time. Without introducing the concept of construction, it was possible to solve this problem, but it looked very strange. A similar problem was solved by me when modeling plans for the production of work, which were simultaneously several versions: optimistic, pessimistic and real. At the same time, the work plan, in turn, was part of another work plan. And there were 5 such floors. Before entering into the model of objects that simulate structures, the simulation looked like this: a lot of links “part-whole”, “colored” in different colors. “Red” links modeled one construction of the object, “green” - another. There were a lot of “flowers” ​​and there was a problem of joining different colors. In fact, these “colors” modeled different points of view on the construction of an object, without naming it explicitly. We had to do the same with the properties of the object to which the properties of the structure were transferred: we had “red” property values ​​and “green”. So we came out of the situation before the introduction of the concept of "construction". I wonder how a similar case is modeled in ISO 15926?
Another practical case: power lines on the one hand, can be divided into routes, each route - into wires. On the other hand, each route can be divided into sections of the route between the supports and so on.
Thus, power lines can be disassembled into parts in different ways. And each method solves a specific practical problem. How, in this case, should the analyst model these constructions, guided by ISO 15926?
There is an interesting method of dividing the same object into parts in different ways. This technique works when the objects into which we divide an object belong to different subject areas. For example, the same object we can call the company, and we can call the function. These are two different paradigms of the presentation of the same. Then we divide the functions separately, the enterprise separately. In principle, if you can add new types of objects, then that part of the problem, which is associated with the modeling of objects in different paradigms, is closed in this way.
Structural modeling using part-to-whole relationships is quite common, because it greatly reduces the size of the model and simplifies the algorithms for working with it. Therefore, often, analysts use this method of modeling. However, this method imposes restrictions on the number of simultaneously existing versions of structures, forces all branches of the enterprise to work with one model of structures, even if this model is counterproductive for someone. At the same time, if we are talking about the construction of objects, then different branches of the enterprise can somehow agree, then when modeling functional structures, such an agreement becomes, in my opinion, impossible. Therefore, returning to the ISO 15926 standard, I am afraid to assume that he was sharpened to model only two points of view on what is happening and what exists. To do this, it has two types of objects: physical and functional. Moreover, each time when modeling two points of view, the modeler has to make a difficult choice between what to call physical and what to call a functional object. Because both constructions can be simultaneously functional, or simultaneously physical objects. For example, is the power line between the supports - is it a physical or functional object? It can be said that it is physical, but if the customer says that the function of this section is energy transfer over a distance, then the section of power lines between the supports will become a functional object, and it will not be possible to simulate two different structures of one power line. Or, a more obvious example: a hydrogen molecule, on the one hand, consists of atoms (one system), and, on the other hand, consists of nuclei and electrons (another system). It is clear that the nature of these systems is the same - physical. How will ISO 15926 model these two different physical structures?
The problem with OOP programming is the same: the construction in OOP is modeled using the aggregation of objects, in fact, part-to-whole relationships. I cannot imagine an object in OOP that can be represented as different constructions. Because the PLO is also sharpened for modeling structures, but only from one point of view. In the PLO, it is impossible to build even two different constructions of one object. How to model the fact that a power transmission line consists of highways and at the same time consists of sections of power lines between supports in an OOP?
A few more words about the place of construction in our thinking, and, therefore, modeling. There are two ways to achieve understanding - synthesis and analysis. When we do an analysis, we see objects as generalized constructions, when synthesis, on the contrary, we represent generalized constructions as objects. Making the analysis, we try to understand how the object is arranged, making the synthesis, we are trying to simplify the model, generalizing it. It turns out the chain: ... an object - its construction - an object (an element of this construction) - a construction of an object - an object (an element of this construction) - a construction of an object ... Next I will not repeat “generalized”, because I will always mean this class of structures. You can start modeling both from the object and from the design. You can move both down and up the hierarchy of objects, making analysis, or synthesis. In another way, this can be represented as approaching or moving away from objects. Approaching, we make the analysis, making the description more detailed, moving away - synthesis, or generalization. Quite funny, but in modern modeling standards I read a lot about decomposition, but very little about composition. If there is something dedicated to the composition, it is written about it with incomprehensible words that are rather difficult to interpret. For example, when we collect statistics on operations in accordance with the methodology of Shewhart, we get the parameters of objects (functions), but the objects themselves are not called. When we model processes and decompose operations, for some reason we cannot do the inverse operation — the composition of processes in an operation. Or the process of describing the subject area is for some reason called “analysis”. But why not "synthesis"? In my opinion, the analyst is engaged in both those and other processes: both synthesis and analysis. Building statistical reports, we are engaged in the synthesis, sorting objects into parts - analysis.
But even with an analysis that seems to be well described in the standards, difficulties arise in the implementation.
To begin to think freely and get away from the limitations imposed by standards, tools and frameworks, I propose to think about why we use constructs in our thinking? As I said, analysis and synthesis can be imagined as a more detailed, or more general description of the subject area. It can be assumed that the design is a tool of such thinking, which allows you to move in the direction of a more detailed, or more general representation of the subject area. If you look at the structure from such an angle, the structure turns out to be a means of modeling in the cycle of synthesis and analysis (the hermeneutic cycle), which we use to achieve understanding. Every time we clarify the details of an object, we actually build its construction. And if so, then we can generalize the concept of construction and imagine a generalized construction consisting of one object. For example, we can say that a city is an object with such coordinates: (...; ...). You can detail this information and say that the city is a square with vertices and list the coordinates of the vertices. Further, it can be said that the city consists of one square-shaped area with the same coordinates. In the new construction paradigm, the first city consists of a second city, which consists of a district. Using the notion that the construction is a more detailed idea of ​​the object, we came to the conclusion that we saw the construction where we would not have thought it before. When simulating maps, I think the task of more detailed representation of objects could have been solved exactly this way.
The design should help us learn something new about the object. For example, consider a flat shape, parts of which will also be flat shapes. The possibility of such a division allows us to introduce the concept of Jordan measure, which, in turn, allows us to introduce the concept of area. Due to the division of the object into similar objects, we were able to introduce the concept of measure. Thus, the division of water into water allows us to learn something new about water - its volume. Therefore, I would also call the division of water into water a construction, and in the definition of construction I sewed up the thesis that it serves as a tool for achieving understanding.
What is missing in the standards of modeling? First of all, the description of the class of tasks that they solve. Standards are good because they allow different subjects to create a model that they understand in the same way as part of the tasks that they solve, to automate the solutions to these tasks, to establish the exchange of information between different information systems within the framework of the tasks being solved, and so on. What is wrong with them? It is bad that the standards are bad, and often the boundaries of their application are often not described. Therefore, the standards, sharpened by the solution of one class of problems, tend to extend to the solution of another class. If the circle of tasks is outlined, then an attempt to solve a problem that goes beyond this circle should lead to a change in the standard, or to abandon it.
The task of creating a unified information model based on a unified ontological basis is becoming popular. At the same time, they often take some kind of industry standard as a basis and try to otmappit the solution of all tasks to this standard (that is, they try to use it as a basis, then to expand it). But this is impossible, if only because different branches of human activity produce the division of objects in a different way. Therefore, adding new knowledge to a unified information model is associated with the creation of new structures and new objects that need to be mapped onto already existing objects and structures.
In the last article I have repeatedly drawn attention to the fact that an object, an operation, a function are the same thing, interpreted by the subject differently. This thesis allowed me to introduce different paradigms of constructions. However, it turned out that the model of activity and the model of objects, as many believe, are completely unmixed models. Let me explain why this is wrong. Suppose we are watching the flow of the river. Our eyes move to the right and to the left, following the slivers and bubbles on the surface of the water. This is the perception of events or operations. After a short observation in this mode, the consciousness gets tired and the eyes stop and the consciousness begins to see something whole - the flow of events. This is a completely different type of perception, corresponding to the perception of the river as a stream, or function. Then the consciousness continues to observe day after day and, sooner or later, ceases to see the flow, and begins to see the object. The river becomes an object devoid of any mobility. Otherwise it can be imagined as if we were moving away from the river. Sooner or later, it would become just an object. In my definition, an object is something that we perceive as unchanging, but that actually changes, since there is nothing unchanging in nature. Another example: drum beats, drum beat rhythm and work. This description of the same, but given with varying degrees of detail. If you were attentive, you noticed that consciousness produces a generalization according to the following scheme: operation, function, object. These three different types of representations of the same, but with varying degrees of detail. You can start with the object and move in the direction of detail. For example, you can start with modeling an enterprise, then simulate a function, and then break it into operations. Thus, the same will be represented first as an object, then as a function, and then as a set of operations. That is, the function is the construction of the object, the operation is the construction of the function.
Design
The explanatory dictionary Ephraim defines two different concepts, which are denoted by one term construction:
- The composition and mutual arrangement of parts of any structure, mechanism.
- The structure itself or mechanism with such a device.
Let's try to translate them into a formal language.
')
Since composition is a set, the first concept is translated as follows: a construction is a set of objects interconnected by connections. At the same time, judging by the definition, the objects must be man-made and inanimate. That is, it is impossible to imagine the Earth in the form of a construction, if you do not assume that it was made by aliens. DNA cannot be represented as a construct, unless this DNA is created by someone. That is, in the definition of the structure it is necessary to add that the objects are man-made. For example, a set of objects: {fuselage, wings, tail} consists of man-made objects, and, therefore, may be called a structure. The design called the plane. I note that in this context, an airplane is not an object, but a multitude of objects (fuselage, wings, tail}. You can call it a multitude of aircraft (k).
How many objects can be in a design? There is no answer to this question in the definition. But we can assume that their final number is greater than one, because the definition speaks of connections. In total, it turned out: the man-made set of objects created by man, the objects are combined by connections, the set is finite, the number of elements is more than one.
In this case there is no obligatory condition for the construction to have a name, or explicitly to indicate the object whose construction is being considered. You can simulate and nameless design.
The second concept of the term “construction” means the following: a construction is an object that can be represented as a set of objects. For example, since an aircraft as an object can be represented as a set of objects consisting of the fuselage, wings and tail, it is also called a structure. In this text, to distinguish the designation of an aircraft as an object from the designation of an aircraft as a set, we can write: an aircraft (o) can be represented as a set of objects — an aircraft (k).
Any object can be divided into parts. We do not know indivisible objects. That is, any object can be called a construction? Not. Because not every man-made object can be divided into man-made parts. For example, casting (pig), being a man-made object, can not be divided into man-made parts. Therefore, the disc can not be called a construction.
Analyzing the term "construction", we found one important feature of the language: the object and its construction are called by the same name. That is, the plane (o) and the plane (k) in everyday life is called by one name: the plane. It is clear that the object and the set of objects are different concepts. In the dictionary of Ephraim, these concepts are different, but in everyday life there is one name, and therefore, people often confuse them and cannot separate these two concepts, denoted by one term. The same problem was in the process approach, in which the concepts of function, functional structure, script, and so on. called one term - process. Because of this, it seemed to many analysts that the function and the script are one and the same.
The confusion, which arises from the fact that two concepts are called in one word, manifests itself in the answer to the following question: what is this or that object? Answers can be divided into two types:
- The first type of response gives the definition of objects: the aircraft is an aircraft heavier than air with a power plant and a wing that creates lift.
- The second type of answers gives the definition of structures: the aircraft - it is the wings, engines, fuselage and tail.
The first type of definition gives the definition of concepts (o), and the second type - definitions of (k).
Another example: a train is a train of railway cars connected between itself and driven by a locomotive or motor car. In this context, the definition of a train (k) is given. It can be said that a train is a long vehicle for transporting passengers or cargo by rail. This is the definition of the train (o). Interestingly, in the dictionaries you can find definitions of both those and other concepts.
In everyday life, we do not notice the difference between such definitions. For example, a group of analysts is shown a production line layout. Each can see completely different pictures. One will see an object called “production line”, the other - a structure having the same name. Since the object and its construction are completely different concepts, they will see completely different things. Until they agree on a single look at this layout, they will talk about different objects. Well, if the context makes them converge on one point of view. However, this does not always happen. The stage at which it turns out the subject of discussion is usually skipped. Because of this, errors in understanding occur. The same problem arises when we want to build an ontological model. For example, if we want to find out the complexity of the aircraft design using the attribute: “the number of aircraft structural elements”, then we need to find an object in the model to which this attribute is attributed. It cannot be attributed to the plane (o), because it is possible to divide the plane into parts in many ways. Therefore, this attribute must be assigned to a set of objects, but not to an object.
System
Let's see how system engineering copes with this terminological paradox. System engineering gives the definition of a system like this:
- Many elements that are in relationships and relationships with each other, which forms a certain integrity.
We see that this definition is close to the definition of a construction in its first sense. That is, a system is a set of objects with properties:
- The set of objects is finite, by the way, more than two?
- There are connections between objects.
- From the set and connections can be synthesized object
The thesis about the man-made system is deleted from the list of properties. The system can be called not only a hand-made product, but also what exists in nature besides our participation. For example, biological systems. Therefore, the concept of a system is wider than the concept of construction. It can be said that constructions are a specialized subclass of the class of systems consisting only of man-made systems.
If for structures the relationship between an object and its structure was called so: “object construction”, then another term is used to designate the relationship between an object and its system: “object structure”. For example, the structure of a person connects a person (o) with a person (s). By the way, I wonder why there is no term “object system” by analogy with the term “object construction”?
Is it possible to call a system an object, and not a multitude of objects? That is, is it possible to apply the term system to an object in the same way as the term construct can be applied to an object? Most likely - you can. For example, it is said that the system has emergence. Formally, this thesis is translated as follows: the properties of an object, the structure of which is represented in the form of the system under study, are different from the properties of elements of this system. Since in this context an object is called a system, an object can also be called a system.
Since any object can be divided into parts, any object can be called a system. This distinguishes the term system from the term construction, because not any object can be called a construction.
It seems to me that in order to eliminate collisions that can arise for an engineer reading books on systems engineering, it is worth making the second definition of the term system into the dictionaries, by analogy with the second meaning of the term construction:
- A system is also called an object whose structure can be represented as a system.
It would be nice if the texts on system engineering clearly indicate what kind of concept in this context refers to an object or a set of objects.
Is it possible to extend to the systems the thesis that any object can have different structures depending on the observer? Yes you can. We are well acquainted with two different paradigms of human structure, which give rise to different structures: the internal structure and external structure of a person.
In system engineering, there is also a requirement that imposes restrictions on the set of possible objects. It's about emergence. An object whose structure is represented as a system must have properties that are different from the properties of the elements of the system. This raises two questions:
- Who determines if a set of objects has an identity or not? For example, let there be pipeline sections. The construction of these sections forms a larger section of the pipeline. Does this site have new properties? If not, then system engineering will not be able to call the construction of pipeline sections a system. If there is a subject who says that there are new properties, then many sites will turn into a system. That is, the subject decides whether a set of objects is a construction or not. Who is he? By the way, a similar problem exists in the definition of a business process in the part where the process, according to the author of the definition, should have a goal. It turns out that the same process, depending on who is looking at it, may or may not be a process.
- The second question is about the docking of system engineering with the ISO 15926 standard. If they are joined, then the question arises of how to model structures in ISO 15926 that do not possess emergence. For example, let a systems engineer say that two pipeline sections combined into one section do not have emergence. However, at the customer such a division of pipeline sections takes place. Systems engineering should either name the sets of objects that are not emergent, and explain how to model them in ISO 15926, or say that system engineering does not consider such sets, and ISO 15926 does not allow them to be modeled.
Returning to the definition of a structure, one can ask the question: should the structure (o) have the properties of emergence? Suppose that should. Then the set of constructions (k) is a subset of the set of systems (c). If not, then the set of constructions (k) intersects with the set of systems (c).
Generalization of the concept of construction
Now let us try to generalize the concept of construction (k) and system (c) to a wider class of objects and sets. In my article I wanted to do just that. Apparently, without the current entry it was not clear. I introduced the concept of a generalized construction (k), which differs from the generally accepted concept of construction as follows:
- A generic construction denotes a set of objects interconnected by connections, but does not denote an object synthesized on this set. This allows me not to specify the prefix (k) after the term construction.
- A generic design may include a set of elements consisting of any number of objects. This means that there can be an empty set, a set consisting of one object, a set consisting of a countable number of objects, a continuum of objects, and so on.
- A set can consist of a set of objects.
- Objects can be of any nature.
- Emergence and other possible criteria are not mandatory conditions for a generalized construction.
In the last article I removed the word generalized from the name of the concept to be more concise.
The result is a hierarchy of classes: A generic structure is the widest set, of which systems and structures are a subset.
It took me to introduce the generalized structure to bring all the structures that we create to describe various structures, as well as to describe the limitations that arise when simplifying these structures, to a single view.
For example, most often, the modeling of structures is made using the “part-whole” links. In this case, information about the design (the average mass of structural elements, for example) we transfer to the model of the object, the design of which we model. The limitations of this modeling method are that we cannot create several different constructions of a single object, whether constructions in different paradigms, or constructions in the same paradigm, but differing in versions.
Once I was tasked with simulating different versions of the design of a single spacecraft. Versions existed simultaneously in time and modeled different versions of design solutions. In addition, the versions themselves changed over time, because design solutions evolved over time. Without introducing the concept of construction, it was possible to solve this problem, but it looked very strange. A similar problem was solved by me when modeling plans for the production of work, which were simultaneously several versions: optimistic, pessimistic and real. At the same time, the work plan, in turn, was part of another work plan. And there were 5 such floors. Before entering into the model of objects that simulate structures, the simulation looked like this: a lot of links “part-whole”, “colored” in different colors. “Red” links modeled one construction of the object, “green” - another. There were a lot of “flowers” ​​and there was a problem of joining different colors. In fact, these “colors” modeled different points of view on the construction of an object, without naming it explicitly. We had to do the same with the properties of the object to which the properties of the structure were transferred: we had “red” property values ​​and “green”. So we came out of the situation before the introduction of the concept of "construction". I wonder how a similar case is modeled in ISO 15926?
Another practical case: power lines on the one hand, can be divided into routes, each route - into wires. On the other hand, each route can be divided into sections of the route between the supports and so on.
Thus, power lines can be disassembled into parts in different ways. And each method solves a specific practical problem. How, in this case, should the analyst model these constructions, guided by ISO 15926?
There is an interesting method of dividing the same object into parts in different ways. This technique works when the objects into which we divide an object belong to different subject areas. For example, the same object we can call the company, and we can call the function. These are two different paradigms of the presentation of the same. Then we divide the functions separately, the enterprise separately. In principle, if you can add new types of objects, then that part of the problem, which is associated with the modeling of objects in different paradigms, is closed in this way.
Structural modeling using part-to-whole relationships is quite common, because it greatly reduces the size of the model and simplifies the algorithms for working with it. Therefore, often, analysts use this method of modeling. However, this method imposes restrictions on the number of simultaneously existing versions of structures, forces all branches of the enterprise to work with one model of structures, even if this model is counterproductive for someone. At the same time, if we are talking about the construction of objects, then different branches of the enterprise can somehow agree, then when modeling functional structures, such an agreement becomes, in my opinion, impossible. Therefore, returning to the ISO 15926 standard, I am afraid to assume that he was sharpened to model only two points of view on what is happening and what exists. To do this, it has two types of objects: physical and functional. Moreover, each time when modeling two points of view, the modeler has to make a difficult choice between what to call physical and what to call a functional object. Because both constructions can be simultaneously functional, or simultaneously physical objects. For example, is the power line between the supports - is it a physical or functional object? It can be said that it is physical, but if the customer says that the function of this section is energy transfer over a distance, then the section of power lines between the supports will become a functional object, and it will not be possible to simulate two different structures of one power line. Or, a more obvious example: a hydrogen molecule, on the one hand, consists of atoms (one system), and, on the other hand, consists of nuclei and electrons (another system). It is clear that the nature of these systems is the same - physical. How will ISO 15926 model these two different physical structures?
The problem with OOP programming is the same: the construction in OOP is modeled using the aggregation of objects, in fact, part-to-whole relationships. I cannot imagine an object in OOP that can be represented as different constructions. Because the PLO is also sharpened for modeling structures, but only from one point of view. In the PLO, it is impossible to build even two different constructions of one object. How to model the fact that a power transmission line consists of highways and at the same time consists of sections of power lines between supports in an OOP?
Place design in the process of thinking
A few more words about the place of construction in our thinking, and, therefore, modeling. There are two ways to achieve understanding - synthesis and analysis. When we do an analysis, we see objects as generalized constructions, when synthesis, on the contrary, we represent generalized constructions as objects. Making the analysis, we try to understand how the object is arranged, making the synthesis, we are trying to simplify the model, generalizing it. It turns out the chain: ... an object - its construction - an object (an element of this construction) - a construction of an object - an object (an element of this construction) - a construction of an object ... Next I will not repeat “generalized”, because I will always mean this class of structures. You can start modeling both from the object and from the design. You can move both down and up the hierarchy of objects, making analysis, or synthesis. In another way, this can be represented as approaching or moving away from objects. Approaching, we make the analysis, making the description more detailed, moving away - synthesis, or generalization. Quite funny, but in modern modeling standards I read a lot about decomposition, but very little about composition. If there is something dedicated to the composition, it is written about it with incomprehensible words that are rather difficult to interpret. For example, when we collect statistics on operations in accordance with the methodology of Shewhart, we get the parameters of objects (functions), but the objects themselves are not called. When we model processes and decompose operations, for some reason we cannot do the inverse operation — the composition of processes in an operation. Or the process of describing the subject area is for some reason called “analysis”. But why not "synthesis"? In my opinion, the analyst is engaged in both those and other processes: both synthesis and analysis. Building statistical reports, we are engaged in the synthesis, sorting objects into parts - analysis.
But even with an analysis that seems to be well described in the standards, difficulties arise in the implementation.
- Often, the types of objects into which (objects, not types) the object being modeled can be divided, are specified rigidly in the standard, code, data structure. For example, in the standard it can be said that power lines are divided into highways. But for other practical tasks (for example, for brigades serving the power lines), this division is not enough. For them, it is necessary that in order to solve some problems of power transmission lines, they are divided into highways, and in order to solve others, they are divided into sections of power transmission lines between supports. If the standard does not allow to do this, then in solving some practical problems one has to manage with crutches.
- The second difficulty is more subtle and less obvious - this is the limit of division. Almost always there are certain atomic objects, on the basis of which the assembly of complex objects takes place. At the same time, it is impossible to divide atomic objects into parts, because the code or data model does not allow this. For the first time I ran into this problem when designing technical accounting for telecom. The twisted pairs were sewn into the code as indivisible entities, for which the code contained an algorithm for finding them as final leaflets in the structure. When it was necessary to simulate a raznazhalka (this is when two wires are taken from different pairs), there were great difficulties with the algorithm for finding pairs. After all, now the twisted pair has ceased to be the final leaf in the structure and the algorithm for its search has changed. The second time I ran into this was when we were modeling the contract. A contract is a model of agreements reached between counterparties. It turned out that the contract is not an indivisible entity. Just as an agreement may consist of agreements, so a contract may consist of contracts. Since the code was tied to entities like “contract” as atomic entities, the need to divide the contract into parts led to the rewriting of a large part of the code. The number of such cases is countless and everyone will find a lot of similar examples in their projects.
To begin to think freely and get away from the limitations imposed by standards, tools and frameworks, I propose to think about why we use constructs in our thinking? As I said, analysis and synthesis can be imagined as a more detailed, or more general description of the subject area. It can be assumed that the design is a tool of such thinking, which allows you to move in the direction of a more detailed, or more general representation of the subject area. If you look at the structure from such an angle, the structure turns out to be a means of modeling in the cycle of synthesis and analysis (the hermeneutic cycle), which we use to achieve understanding. Every time we clarify the details of an object, we actually build its construction. And if so, then we can generalize the concept of construction and imagine a generalized construction consisting of one object. For example, we can say that a city is an object with such coordinates: (...; ...). You can detail this information and say that the city is a square with vertices and list the coordinates of the vertices. Further, it can be said that the city consists of one square-shaped area with the same coordinates. In the new construction paradigm, the first city consists of a second city, which consists of a district. Using the notion that the construction is a more detailed idea of ​​the object, we came to the conclusion that we saw the construction where we would not have thought it before. When simulating maps, I think the task of more detailed representation of objects could have been solved exactly this way.
The design should help us learn something new about the object. For example, consider a flat shape, parts of which will also be flat shapes. The possibility of such a division allows us to introduce the concept of Jordan measure, which, in turn, allows us to introduce the concept of area. Due to the division of the object into similar objects, we were able to introduce the concept of measure. Thus, the division of water into water allows us to learn something new about water - its volume. Therefore, I would also call the division of water into water a construction, and in the definition of construction I sewed up the thesis that it serves as a tool for achieving understanding.
Limitations of modeling standards
What is missing in the standards of modeling? First of all, the description of the class of tasks that they solve. Standards are good because they allow different subjects to create a model that they understand in the same way as part of the tasks that they solve, to automate the solutions to these tasks, to establish the exchange of information between different information systems within the framework of the tasks being solved, and so on. What is wrong with them? It is bad that the standards are bad, and often the boundaries of their application are often not described. Therefore, the standards, sharpened by the solution of one class of problems, tend to extend to the solution of another class. If the circle of tasks is outlined, then an attempt to solve a problem that goes beyond this circle should lead to a change in the standard, or to abandon it.
The task of creating a unified information model based on a unified ontological basis is becoming popular. At the same time, they often take some kind of industry standard as a basis and try to otmappit the solution of all tasks to this standard (that is, they try to use it as a basis, then to expand it). But this is impossible, if only because different branches of human activity produce the division of objects in a different way. Therefore, adding new knowledge to a unified information model is associated with the creation of new structures and new objects that need to be mapped onto already existing objects and structures.
The unity of the object, function and operation.
In the last article I have repeatedly drawn attention to the fact that an object, an operation, a function are the same thing, interpreted by the subject differently. This thesis allowed me to introduce different paradigms of constructions. However, it turned out that the model of activity and the model of objects, as many believe, are completely unmixed models. Let me explain why this is wrong. Suppose we are watching the flow of the river. Our eyes move to the right and to the left, following the slivers and bubbles on the surface of the water. This is the perception of events or operations. After a short observation in this mode, the consciousness gets tired and the eyes stop and the consciousness begins to see something whole - the flow of events. This is a completely different type of perception, corresponding to the perception of the river as a stream, or function. Then the consciousness continues to observe day after day and, sooner or later, ceases to see the flow, and begins to see the object. The river becomes an object devoid of any mobility. Otherwise it can be imagined as if we were moving away from the river. Sooner or later, it would become just an object. In my definition, an object is something that we perceive as unchanging, but that actually changes, since there is nothing unchanging in nature. Another example: drum beats, drum beat rhythm and work. This description of the same, but given with varying degrees of detail. If you were attentive, you noticed that consciousness produces a generalization according to the following scheme: operation, function, object. These three different types of representations of the same, but with varying degrees of detail. You can start with the object and move in the direction of detail. For example, you can start with modeling an enterprise, then simulate a function, and then break it into operations. Thus, the same will be represented first as an object, then as a function, and then as a set of operations. That is, the function is the construction of the object, the operation is the construction of the function.
Source: https://habr.com/ru/post/328192/
All Articles