The primary source of related data in the design of databases of any kind is the corresponding subject area , which will be considered as a collection of knowledge and data about objects and processes , to be designed and stored in DB. In this case, the problem of imitating modeling of complex computer systems is considered as a subject domain.
Knowledge of the subject area, as noted in § 3.1, devoted to the methodology for the development and machine implementation of the .U system models, can be obtained from a variety of sources, such as:
- Fundamental laws and stable laws, theoretical knowledge of applied sciences of processes in the simulated system 5;
- heuristic knowledge in the form of accepted rules, conventions and notations, characteristic for systems 5;
- Expert knowledge and expert assessments of specialists in the field of modeling specific systems 5;
- existing databases, computerized systems, technologies and projects.
The listed sources of knowledge about the subject area "system modeling" are not the sources of the actual data, but determine the methodological principles for the allocation of application objects and processes, as well as methods for acquiring and binding data on the selected objects and processes. Sources of knowledge about the subject area determine the choice of the classification scheme of the formalized description of objects and processes, which in the course of database design is displayed in the resulting relational schema.
Thus, database design process according to the complete technological scheme is the process of step-by-step mapping of the initial classification scheme of the domain into the relational database implementation scheme. The choice of this or that methodological basis of the database design process determines the type of relationships between the data implemented in the database and characterizes the integration capabilities of the built database.
An important methodological feature of the projected database and its integration capabilities is the natural or artificial nature of the allocation of applied objects and processes, the associated data on which will be stored in the projected database. Of particular importance is this circumstance when using the object-oriented approach to the design of databases and the construction of IRDD.
In the practice of modeling complex systems 5, when designing local databases (LBDs), the application of classification schemes is often performed intuitively, with the firm conviction that the chosen way to determine the data of the database is natural for solving this problem. In fact, the LJD designer simply accepts one of the existing classification schemes (for example, the traditional database construction scheme "system 5 - the conceptual model M to - the machine model M m from the literature). At the same time, the lack of knowledge of the rules for the formation and binding of attribute values even in such a database can lead to distortion and errors in the interpretation of data extracted from the database as information.
Conceptual analysis and database design.
The choice or construction of a classification scheme for the selection of objects and processes is the essence of the analysis of the subject area (in this case it is the "modeling of complex systems 5"). The mapping of the classification scheme into the resulting relational schema defines a multi-step process for designing a PM.
First of all, the formalized representation of the domain will be understood as the classification selection of objects and processes, which in terms of the object-oriented approach corresponds to the classification definition of the abstractions of essences and behavior [2,16]. Classification can be performed by classical categorization, conceptual clustering, or using prototype theory methods. Entity abstractions define classes of objects of the domain and are expressed through sets of characteristic properties of objects of a given class, at the same time, distinguishing objects of this class from objects of other classes. Behavioral abstractions express rules for the interaction of objects through common or related properties.
Thus, the chosen or constructed classification scheme with application of this or that classification methodology determines the system of formation of the universe of the projected database. Fundamentally important for database technology is the principle of separating the classification scheme in the form of a set of elements for modeling data representations from the actual system of classified database data. It is the implementation of this principle that can lead to the resolution of the problem of constructing a certain category of object-oriented databases. Elements of modeling data representations are also included in the universe.
Designing a conceptual representation of a system of classifying a domain based on the chosen classification scheme is called conceptual design of the subject area, as a result of which a conceptual model of a family of conceptually homogeneous databases is constructed. Obviously, the choice of ready-made design solutions refers the projected database to a database family with a homogeneous conceptual representation. For example, the typification of the classification of data about the model of system 5 (the conceptual model, the typical mathematical scheme - see Chapter 3) practically excludes the need for a conceptual design of such a fragment of the domain. All databases, including descriptions of models of system 5, according to the accepted classification scheme for determining the mathematical scheme, in this fragment the databases are conceptually homogeneous. Note that such a classification scheme has an artificial nature of construction, based on a heuristic system.
The situation is completely different if a database fragment is designed, describing the properties of material objects (for example, modeling objects, ie, systems 5). The classification scheme for this case will be based on the methodology of classical categorization, and knowledge of the relevant subject area will be fundamental. In this case, the conceptual design will be associated with the need for serious scientific and methodological work on abstracting the knowledge system, usually described in dictionaries, handbooks, thesauri. Achieving the conceptual homogeneity, and, consequently, sufficient integration capabilities of a family of similar databases becomes an extremely complex problem. Thus, the conceptual modeling of the domain includes the following operations:
- determination of the nature of the source of knowledge about the subject area (fundamental, heuristic, expert, existing computerized solution in the field of modeling specific systems 5);
- selection or construction of a classification scheme based on the classification methodology (classical categorization, conceptual clustering, prototype theory);
- highlighting the abstractions of objects and processes (the interaction of objects) through the definition of their properties;
- construction of the structure (hierarchy) of the base classes of the domain and the formalization of the representation of such a structure.
As a result of conceptual modeling, the following conceptual components are defined:
- a set of typed abstract representations of entities of the domain, defining a system of rules and restrictions on the formation of attribute values of the properties of objects;
- a set of typed abstract representations of the interaction of entities of the domain;
Classification rules for the formation of attributive values of the properties of objects, as well as rules for the interaction of objects through their specific properties form the conceptual semantics of the subject area in the basis of the chosen classification scheme. Abstractions as a whole and their separate properties, isolated on the basis of fundamental knowledge of the domain, form a fundamental semantics, whose scope of action goes beyond the boundaries of the given subject area. Fundamental semantics is the basis for global integration of the database by definition. Conceptual semantics is the basis for building prototypes of indexing schemes and setting constraints for elements of relational database schemas. In other words, building or owning conceptual semantics predetermines the ability to extract information from an implemented database. To a large extent, the conceptual semantics governs the procedures for normalizing and ordering relational schemas.
The conceptual model of the domain is the basis for the design (and integration) of a family of conceptually homogeneous PM. The presence of fundamental semantics can essentially increase the integration capabilities of the corresponding PM. The conceptual model forms the header sections of the universe, defines the interface of a family of conceptually homogeneous databases, expresses a number of conceptual components (abstractions, connection properties), on the basis of which it is possible to implement variants of infologic design by combining databases.