138. Excitation (Fire) excited a conversion. See Run To, OMPLETION, Trigger. Semantics When the event of the conversion requirement, if the monitoring condition is met, the conversion will perform its activity, the active status changes. After the object is received, the event is saved if the state machine is running to the completion step. The step is processed by the state machine after the step is completed. If the status of the current object contains a conversion, the corresponding conversion is triggered. If there is a complex conversion of multiple source states, the conversion must be active before the transformation is performed. After the source status activity is complete, you can perform the completion conversion. If the source status is constituent, all direct subsystems are completed or reached after the final state can perform complete conversion. After processing the event, calculate the conditioning condition (if any). If the Boolean expression of the condition is true, the conversion will be excited. The activity on the transition is excited, and the target state of the conversion is active (internal conversion does not change). In status changes, any exit activities and entrance activities on the original state of the object and the transition target state shortest path are performed. Note that the original state may be a nested state of the converted source state. If the conversion does not satisfy the guard condition, it will not be stimulated. There may be a conversion that may arouse other guardians. If there are multiple conversions, it can be excited, only one of them can truly stimulate. The conversion in the nested state is prioritized than external conversion. Otherwise the choice of conversion is uncertain, which is in line with common situations in the real world. In fact, implementation can determine the order in which the conversion excitation is determined. This does not change semantics. Reasonable setting of monitoring conditions to ensure mutual exclusion, can achieve the same effect. A simpler method is to use "If other conversions are not converted, use this conversion". Deferred Events If an event is delayed in a certain state, it will not trigger the conversion until the object enters a state location that does not require the current event delay. Once the object enters a new state, the original delay and now the event that is not delayed is "Pending" "and appears in an uncertain order. If the first time to resolve the event, it will be ignored, the next, the event is resolved. If the event causes a state transition, according to the delay in the new state, re-determine the delay event that is still in-to-resolution, and establish a new event set to resolve. Implementation can be more stringent to specify the processing order of the delay event, or to limit the order in a certain operation. 139. Flow (FLOW) The relationship between two versions of the same object on different consecutive time points. See Become, copy (COPY). The semantic flow relationship links two versions of the same object in different consecutive time points. It can connect two values of an object in an instance stage, or describe the two type element characters of the same object in interaction. It represents the object from one state to another, represents the value, the control state, and the location change. The structure type of stream dependencies is becoming (Become) and replication (COPY), and users can add other constructive types. Indicates the method of constructing the type of dashed arrow with the structure type, and the structure type cannot be omitted. Standard elements become, copy 140. A Symbol in the Focus Of Control sequence diagram, indicating that the object is directly or indirect (through subroutine), performing a period of time. See Activation. 141. Font USUGE Different fonts or other graphics flags are distinguished. See graphic marker. Discussion of oblique characters represent abstract class, attribute, and operation. Other fonts are used to emphasize or distinguish between different parts of the syntax. It is recommended to use the name of the name of the name; for content, attributes, operations, role names, normal fonts. The name of the point uses special fonts, the option is the editing tool. The tool can select font to emphasize elements, distinguish the reserved word, indicating a specific attribute of the element. You can also allow users to select fonts.
For similar considerations, colored blind people can distinguish between colors. All of the above is an extension of the representation of this book. 142. In complex conversion, one source state can be transferred to multiple target status, enabling the number of active states. Antonym: binding. See complex transformation, composite state, combine (JOIN). Semantic bifurcation refers to a source state and a plurality of target status. If there is a trigger event when active activity is active, all target status will become active. The target status must be different regions in parallel composition. Indicates that the method is expressed as a thick line that is transferred to the arrow and a plurality of transfer arrows, which can have a conversion tag (monitoring conditions, trigger events, and activities). Figure 13-101 is a significant bifurcation of the in parallel composition state. Figure 13-101 Fork 143. Formal Argument See parameters 144. Frameter) 144 provides an extensible template for an extensible template for applications in a domain. See the package. 145. A friend uses dependencies that allow users to access service providers. See Access (Acess), Import, Visibility Semantic Professional Dependence is used to ensure a license for an operation or a class to the content of the class, although it is not sufficient from another perspective. It is usually an exception in the rules, which is careful. Representation is a dashed line representing a dashed line from a class that is licensed or operated to provide content; the structure type keyword "Friend" is marked on the arrow. 146. Full Descriptor A complete implicit description of a direct instance. A complete descriptor is assembled by inheriting all its ancestors. See direct class, inherit, multiple yuan. Semanticity is actually, for the declaration of class or other elements, only part of the examples of the same example, called class segment. Typical objects contain more structures than its direct classes. The description of all the properties, operations, and associations of the example is called a complete descriptor, which is usually not represented by the model or program. The purpose of the inheritance rule is to provide a method of automatically become a complete descriptor in the class fragment group. Theoretically different implementations, called metaobject protocol. UML is a series of rules defined by inheritance, covering most programming languages, and can also be used to establish a conceptual model. Note that there may be other possibilities, such as CLOS languages, etc. 147. Functional View This view will decompose system to function or provide functions. It is usually considered that the functional view is not object-oriented, which may result in a structure that is not easy to maintain. In the traditional development method, the data flow diagram is the core of the functional view. UML does not directly support this view, but there are some functional characteristics in the event map. 148. Generalizable Element can participate in the model elements of the generalization relationship. See generization, inheritance semantic generalization elements can have a parent. Variables that are classified into elements can be used as an example of the descendants of the element. The generalized elements include classes, use cases, other types, associations, status, and cooperation, which inherit its ancestor features. Which part of each perfinitective element is inheritance, to see the type of element. Position and constraints, methods, operations, and associations; associated inheritance participation (itself can be special) and the characteristics of the associated end; use case inheritance attributes, operation, association with participants, and other use cases Contains relationships, behavioral sequences. State inheritance conversion. See generization, associated genreization, use case generalization. The properties of the structural generalized elements declare where it can appear in the generalization relationship. Abstract Description The generalized element is a description of the direct instance or abstract elements.
TRUE indicates that the element is abstract (no direct instance); false means it is a body (can be directly instance). Abstract elements entssure can be used. A class with unparalleled operation is abstract. Ye explains whether the generalized element can be special. TRUE indicates that the element cannot be post-era (leaf); the False table can be proceed (whether or not there is a future generation). As the abstraction class of the leaves, it can only work with global properties and operations. The root indicates whether the element must be the root without ancestor. TRUE indicates that the element must be root; the False table does not have to be root, and there is ancestor (whether there is a ancestor). Note: The declaration of the leaves or roots do not affect semantics, but this statement can give designers. If you can avoid the analysis of global variables or global protection for polymorphism, you can have higher compilation efficiency. Standard Elements Leaves 149. Generalization a wider-gear relationship between a more widely elements and a particular element. The special elements completely include a widely widened element and contain more information. Examples of specialized elements can be used in any place to use widely used elements. See Uliting Generalization, inheritance, Substitutability PrinciPle, Case Generalization. Semantic flueraits are direct relationships between two similar elements. One of the elements is called a parent, and the other is a child. For classes, the father is called superclass, and the child is a subclass. The direct instance of the parent has a common feature of all children. The example explained by the child is the subset of the above instance, which not only has the characteristics of the parent, but also a unique feature. Ubging is an opposite. According to one direction, it is converted into a parent, and the other direction leads to the sub. Elements passing through one or several generalization relationships in the parent direction refer to the ancestors; elements of one or several generalization relationships in the sub-direction are called the child. Unmissions are not allowed, a class cannot be both his ancestors and their own progeny. In the simplest case, classes (or other generalized elements) have a single parent. In complex situations, there are multiple parents. The child inherits all the parents, behaviors, and constraints, called multiple inheritances (or multiple floods). The parent element has visibility to the child element. Associate, class, state, events, and cooperation can be generally described. See the associated generalization for the application of the generalization of the associated. For the prevention of the generalization of the use case, see example extensive. Nodes and components are similar to classes, and their generalization is similar. Constraint constraints can be applied to a column of generalization, as well as the same parent. It can be specified that the following properties: mutually exclusive a ancestor can not have two children (multiple inheritance), the instance cannot be two indirect instances (multiple class symbols) overlap a ancestor can have two or more children, the instance can belong Two or more children. Complete list all possible sons can not be added. Incomplete not listing all possible sons, some known children have no statements, but also add new children. The generalization relationship between the method is expressed as a solid line between the child element (e.g., subclass) to the parent element, the path pointing to one end of the broader element with hollow triangles (Figs. 13-102). Personal lines can be made or trees (Figures 13-102). Figure 13-102 Generalization Relationship Figure 13-103 Tree generalization relationship is also applicable to association, but too many lines may make the picture very messy. In order to indicate the generalized arrow, the association can be expressed as an associated class. If the class exists in the model is not marked, the omitted number (...) should be replaced in the corresponding position. (This does not mean that the new class is to be added in the future, but the current existence of the presentation is existed. This piece is ignored, not a semantic element).
A subclass of a class is expressed as omitted number indicating that at least one subclass is not in the view. The omitted number can have a descriptor. This indicator is automatically maintained by the editing tool without manual input. A set of generalized paths indicating options to a superclass can be represented as a common road (including triangle) tree, branches point to subclasses. This is just a representation, does not represent N-dimensional relationship. In the current model, there is a generalization relationship between the superclass-subclass. The arc has no semantic difference. If you have a text tag on a plurality of subclasses, the label applies to all paths. In other words, it is the properties shared by all subclavants. Examples Examples 13-104 are constraints declared on the generalization relationship. The "tree" generalization is shown (the path is parallel, there is a common arrow); the "binary" generalization is also expressed (each pair of parent-sub-relations have independent arrows). Figure 13-104 Using the elements in the generalization of the generalization relationship without understanding this element, the child element must know the structure of the parent element. Typically, the parent element is designed to provide a child element extension, so it is designed to understand the expected child elements. One of the advantages of generalization relationships is that the child elements that often don't think of the design of the parent element, which brings a stronger function. Implementation relationship is similar to the generalization relationship, but only inherits the behavioral description. If the desired element is not attribute, it is associated, and only abstract operations, the generalization and implementation are the same. Implementation relationship does not generate a user, so the operation must be in the user or from other elements. Standard elements complete, mutually exclusive, implementation, incomplete, overlapping 150. Graphic Marker, a tag element, such as geometry, texture, filling mode, font, color, and the like. The font usage representation indicates that the symbol is made up of different graphic tags. A graphic marker itself has no semantic role, but the purpose of the representation is to form a graphic tag as much as possible. Some graphics markers are used to construct a predefined UML symbol, while others are not used to represent. The color is not given the meaning, because some printers do not support color, some people can't distinguish color. These undefined graphical tags can be used in accordance with the designer or tool requirements. The UML representation only allows limited graphical extensions. Icon or graphic tags (such as textures, colors) can be associated with constructive types. UML does not define the format of a graphical description, but there are many bitmaps and formats available for graphics editors (although their portability is difficult). The general form of icon instructions and its alternative is easy to understand, and we also allow tool developers to dissolve into their own imagination - note that excessive use of scalability may result in weakening of tool portability. 151. Guard Condition must satisfy a condition before the transformation. See branch, Conditional Thread, Junction State, Transition Semantic Watch is a Boolean expression, which is part of the conversion description. After receiving the transformed trigger event, the system saves the event until the state machine has completed any running to the completion step, then the event processing is subsequently processed, and the care condition is calculated. If the condition is true, it can be converted (if there is a plurality of conversions, only one can be performed). Test when the event is processed. If the condition is false, the monitoring conditions are no longer recalculated unless the trigger event appears again. The guard condition must be an inquiry - it cannot modify the value of the system or the value of the status, and there is no side effect. Completion conversion can also have a guard condition, at this time, it selects a branch execution. The method of monitoring the method is part of the conversion string, which is a Boolean expression [Boolean expression] expressed in square brackets [Boolean expression] The name used in the expression must be internal visible, which can be the parameter of the trigger event or the current object Attributes.
152. Guillemets ("") represents references in French, Italian and Spanish. The keyword and constructive type are represented in the UML representation. Many fonts have, if necessary, use two angle brackets. (<< >>). See the font. 153. History State, a pseudo-state, indicating that the current internal composition state remembers the active sub-status of it exists. See the constituent state, pseudo-state, state machine, and the transition semantic historical state allows sequential constituting state to remember the last active sub-state that constitutes a state before the conversion emitted from the constituent state. One conversion of the steering history will make the previous active sub-status becomes active and perform corresponding entry activities and export activities. The historical state can have a conversion from an external state or an initial state. It can have an excessive conversion without a label, which indicates the historical state that is previously saved. When there is no saved state, the steering history will turn to it. Historical states cannot have other conversions from internal components. Historical states can memorize history and deep history. Shallow historical state saves and activates the status of historical state in the same nested level. If a conversion exits the constituent state from the nested sub-state, the terminator state in the constituent state is activated. A deeper nesting level in the deep historical state memory composition. To remember the deep state, the conversion must be transferred directly from the deep state. If a conversion is converted from a deep state to a shallow state, and thereby rotates the constituent state, it will be a conversion of the shallow state. To a new historical conversion recovery any status activated at any level. In this process, if the entrance activity occurs on the inner layer state containing the state of the remembered state, the entry activity is performed. The constituent state may have two historical states at the same time, and the incoming conversion must be connected to one of them. If the constituent state enters the final state, it will discard all saved historical states, as if never enter this state. Indicates that the shallow historical state is represented by a small circle with H, as shown in Figures 13-105. The deep historical state is a circle with H *. Figure 13-105 Historical Status 154. HyperLink two indicators between elements can pass through certain commands. See the picture (Diagram). Indicates that the representation on the paper surface does not contain hidden information. However, the representations on the computer screen can have additional invisible hyperlinks, which do not appear in the static diagram, but can be activated on the dynamic diagram to access other information. This information is or in a graphical view or in a text table. This dynamic link is as part of a dynamic representation like a visual information. This article does not limit their use, but we didn't treat them as a necessary function. This article tries to define a static representation for UML, but some useful information is difficult to represent in these views. On the other hand, we have no understanding of dynamic tools, and we don't want to do word only for innovative dynamic representations. In the end, mature dynamic representations will become standard, but it is currently a fashion. 155. Identity An object's inheritance property is used to distinguish it from other objects. See Data Value, object (Object). Semantic objects are different from each other. An identifier of an object is its concept handle, is a inheritance feature of other object references or determined. Conceptually, objects do not need to be calibrated with primary key (key) or other mechanisms, which should not exist in the model. In implementation, the identifier is implemented by an address or primary key, but they are part of the basic implementation structure without displaying the properties of most models. 156. The structural design of the unnight structure (Ill Formed model is incorrect, with one or more predefined rules or constraints being violated. Contrast: Good structure. See conflict (Constraint). Semantic violations of good structural rules and constraints cannot be used because its semantics are contradictory. Use this model will result in meaningless results. Modeling tools have responsibility to identify non-rootic structures and hidden to prevent even more problems. Since the UML built-in semantic extension is sometimes used, automatic recognition is not always guaranteed.
In addition, automatic check cannot guarantee the consistency of operation. Therefore, in practice, it should be combined with automatic identification and manual identification. The completed model must be good structure, but the model version of a model built in the model can be non-benign because it is just an unfinished framework. Modifying an available model to a new model, or a model of a non-rootic structure may also be generated. This is similar to writing programs - the program for compilation must be available, but the intermediate version of the text editing area can be unused. Therefore, support tools must edit and save models of non-rooted structures. 157. Implementation 1. Define how a thing is constructed, calculated. For example, the class is an implementation of type; method is the implementation of operation. Contrast: Description. It is realized between implementation and instructions. See a Realization 2. Use executable media (such as programming language, database, digitized hardware) to describe system functions. For implementation, the lower level of decision must be produced to make the design suitable for specific implementation media and adapt to the environment (each language has its own limits). If the design is good, the implementation of the decision will be a local area, and any decision does not affect the overall system of the system. This step is captured by the realization layer model, especially static diagrams and code. Comparison: analysis, design, implementation, and configuration. See the development process, modeling steps (stages of model). 158. Implementation Class's physical implementation type, including properties, associated and operation of other classes. One implementation class will be used for traditional-oriented language-oriented languages with static single-category. In this way, an object in the system must have only one implementation class as its direct class. Unlike types, the structure type of the class allows multiple types of elements. In some languages (such as Java), objects can have an implementation class and multiple types, and implement classes are consistent with the type. See the type (TYPE), which is compared to the type and implementation class. 159. Implementation inheritance inheritance completion - inheritance structure (such as attributes and operations) and code (such as methods). The opposite interface inherits includes inheritance (operation) of the interface, but does not inherit the method and data (attributes and association). The meaning of UML in the UML includes the inheritance of the interface and implementation. If you only inherit your implementation (private inheritance), you can use the keyword "importation" on the general relationship. If you use it as long as the interface is inherited, you can use the implementation relationship of the interface. See generization, inheritance, interface inheritance, private inheritance. 160. A view of the Implementation View model, contains static declarations, dependencies, and classes that may be implemented by components. See the component map (Component Diagram). 161. Import (Import) license dependencies, wherein the element name in the provider package is joined to the namespace of the user bag. See Access (Acup), Package, Visibility. The namespace in the namespace of the semantic provider package is added to the namespace of the user bag, and the specified visibility rules will not change. If the imported name and the names of the original namespace conflict, the model is a non-rootic structure. See Access (Acess), as well as visibility rules for access and import. Representation The packet of the access authority is directed to the package of the provider with a dashed arrow, and the arrow has a constructive type "import". 162. Inactive is inactive, it cannot be processed by any object.
163. Inception software development process, this period is designed to design the original program of the system, develop certain analysis views and a small number of other views. See the Development Process. 164. The relationship between the included case and the included case is included. NOTE How to insert a behavior defined by the behavior defined in the use case into a reference case. Base use cases can be seen including the use case and rely on execution results including the use case. But both cannot access the other party's attributes. See Extensions, Use Case, Contains a Semantic Conference Contains Relationship with Alternatives and Contains Examples. The included use case in the relationship is not a class element that can be implemented independently, but an explicitly explaining an additional behavior sequence in the use case example of performing a base case. A basis can have multiple inclusion relationships. The same includes use cases may be included in a plurality of base cases, there is no need to have any relationships between these basis. As long as each inserted in different locations in the alternative example, there may be a plurality of related relationships between the same pairs and including the case. Contains the use case to access the properties or operation of the base case. The use case provides a unique behavior that can be reused by multiple base cases. Base cases can see the included examples of the attribute value, but cannot access the properties of the use case, because the current case is re-controlled, the included case is over. Note: (All) Additional content can be nested. A group containing various, extended or generalized relationships. The structure includes the relationship with the following attribute: the position of the behavioral sequence body of the location base case, and can be inserted in this location. When performing a base example instance reaches this location, the use case instance executes the included case, then continue to perform the reference case. It is included as an explicit statement in the alone, and therefore, the location is implicit. This is different from the extension relationship. The included effects can be reached by reference only to execute only once, by reference to the cycles in the sequence in the sequence of the included alone. Indicates that the broken line arrow is pointed out from the reference case to the included case. There is a keyword "incrude" on the arrow (Figure 13-106). The position can be used as a property labeled on the edge of the arrow, but is usually referred to as a text portion of a reference case, not shown in the figure. Figures 13-107 are examples of examples of examples. Figures 13-106 Contains behavioral sequence 165 of the relationship map 13-107 165. Incremental development system model and other product development form a series of versions, each version is completed to a certain degree of functionality and refinement, Each version adds detailed content than the previous. The advantage of this method is that each new version is a small amount of change in the previous version, it is not easy to go wrong. This approach is closely related to iterative development concepts. See the Development Process. 166. Indirect instance (Indirect Instance) an entity that is an example of an element (such as class), while it is an example of the element. That is to say it is an example, but it is not a direct instance. 167. Inheritance enables a more specific element to comply with a broader element definition structure and behavior in this mechanism. See the full descriptor (Full Descriotor), genia. Semantic inheritance enables the full descriptor of the generalized element to automatically construct it through a declaration section of the assembled level. Ultrafining inheritance is a tree that makes a model element (class) declared (actually a predecessor tree). However, each statement is not for a completely available element, each declaration only shows that the element is more than its ancestor, inherits, integrating these increments as a complete description of the example (implicit) process, these Complete explanation describes the actual instance. It can be considered that each permutable element has two descriptors: a fragment descriptor and a full descriptor. The clip descriptor is an incremental feature table (, for example, the attributes and operations of the class declarations in the model. The elements and their parent fragment descriptors are different, and the full descriptor will not appear in the model. It is an element instance's full description - such as all attributes and operations of the objects. The full descriptor is a consortium of the elements and their parent segment descriptors.
Inheritance is the incremental definition of the element, other details such as finding algorithms, etc., is just a specific language implementation mechanism, not part of the core definition. Although this explanation seems to be a bit strange, it does not need to implement the implementation process in most other definitions, while compatible with these definitions. Conflicts If the same feature occurs many times in the inherited clip collection, a conflict may occur. The attributes in the inheritance focus can only be declared. If a few declarations will result in a conflict, the model is a non-rootic structure. (This limitation is not based on logical reasoning, but to prevent paramutation that may generate when there must be pathname properties, can be declared multiple declarations, as long as the statement is the same (the method can be different), or the sub-declaration is enhanced Inheritance declaration (such as a sub-declaration as a queue or increasing its parallel status). The method of sub-declaration replaces (overloaded) ancestor statement, no conflict. If a different method is inherited from two elements without ancestral relationship, a conflict occurs, and the model is a non-alarming structure. Discussing the generalization of the elements, which explains what is an element; inheritance is a mechanism that connects the sharing increment declaration, thereby forming a complete statement of elements, which is different, but has a close relationship. Applying the inheritance mechanism in the generalization relationship, you can implement the decomposition and sharing, and polymorphism. This is the way UML and most object-oriented languages, but some programming languages may also use other methods. 168. The target state of the initial state converted is the boundary of the constituent state to indicate the pseudo-state of its default starting position. See Composite State, Creation, Initialization, Junction State Semantic Semantic Semantic Semonics should be separated from internal details of the external view as much as possible . From the outside, the constituent state is the opaque entity that hides the internal structure. From the external view, the stop point of the conversion is the state itself. From the internal view, they are connected to the status of the status. The initial state is a pseudo-state that represents the joint point of the entry conversion and the constituent state boundary. It is not a real state, can not reserve control, it is just a grammar expression that controls should be turned to. The initial state must have an outgoing no contact (there is no transition to the event trigger, so encountering the initial state can be automatically valid). Complete the conversion to a true state in the constituent state. The completion of the conversion can have actions, which is executed when the inlet action (if any) is encountered. In addition to the entrance activity (executed at all portions, it will also allow the action to be associated with the default inlet. This activity can access the incident implicit current event - that is, events triggered by the first segment of the transition that transitions to the initial state. The initial state cannot have an outward conversion of triggering events, and entering conversion is the same as the conversion of entering its constituent state, and should be avoided, such conversion should be connected to the constituent state. Usually the transition on the initial state is no monitoring condition, so it must be unique conversion from the initial state. Multiple outgoing conversions will have a monitoring condition that must cover all possibilities (or set a certain condition to "Other"). The key is that control must leave the initial state immediately. It is not a real state, you must activate a conversion. The initial state in the top state of the class represents the creation of a new instance of the class. When it performs its excessive conversion, an implicit current event is an event for creating an object, which has a variable value from the configuration operation. These values can be used for activities on the external conversion. Object Creating a class of initial status of most of the components of a class is similar, which may have a trigger event with keyword "create", as well as a nameful trigger event with parameters. There are multiple such conversions having different trigger events.
When a class object instance is generated, the conversion corresponding to the creation operation is triggered, which obtains a variable value from the call creation operation. The initial state in the representation state is denoted as a small solid circle, which can except the arrows thereon. There can only be an initial state (directly) in a constituent state. However, there may be multiple constituent states in the nested composition state. For example, in Figures 13-108, we started from the state x, after the event E, the transition and execution of the active A, the conversion pointing state y, after performing the inlet activity B, the initial state is active. Expendent conversion is then performed immediately, and the active C is executed and the state Z is entered. If the event f occurs when the system is in state x, activate another conversion execution activity D, the conversion directly reaches the status z, which does not include the initial state during this process, because the control is translated to Y, so do not execute B, but not executed Activity C. The initial state in Figures 13-109 has a branch. Suppose the system begins in the x state. When the event e appears, the system is executed, the system enters the status y and completes the entry activity B, the control is transferred to the initial state, and the size attribute of the current object; if the value is 0, control transfer Status z; if the value is not 0, control the transfer status W. Figure 13-108 Initial State Figure 13-109 Initial Value, an initial value, an initial value, an initialization, an initial value, an initial value, an initial value, a default value, an initialization. Semantic initial values are an expression that attribute belongs. Expression is a text string that can be translated in some language. When an object with this property is initialized, the expression is calculated using a given language and system current value. The calculation results are used to initialize this property in the new object. The initial value is optional. If not, the attribute declaration does not indicate the value of the newborn object (the other part of the model will provide this value) Note that the explicit object initialization process (such as constructing) can ignore the initial value expression Or overlay attribute values. The initial value of class range properties is only used at the beginning of the execution, and UML does not specify a relative order of the initialization of different class attributes. 170. Initialization assigns a newly generated object - that is, its properties, connection and control status of the belonging. Semantic concept, the creation of new objects is completed, but it is more easily understood by two steps: Creating and initialization. The empty shell object is given, which has an appropriate structure and attribute location, and imparts the identifier, the identifier can be implemented in different ways, such as the memory block address, or internal counter containing the object. In summary, the identifier is unique in the system, which can be used as a handle for finding and accessing an object. To this end, the object is still not legal - it violates its values and constraints. The next step is to initialize, the given initial value expression, add the result to the corresponding attribute location. Creating methods can explicitly calculate attribute values to overload the default initial expression, and the result value must meet the constraints of the class. Creating methods can also generate links to new objects that must meet the requirements of associated multiplex associated with classes. Once the initialization is completed, the object is legal and should comply with all the constraints of its class. After the initialization is completed, the attribute or its variability attribute is the association of Frozen or Add Only to change until the object is destroyed. This initialization operation is atomic, and cannot be interrupted, and it cannot be left in the middle. 171. Instance (instance) with an identifier and a separate entity. The descriptor explains the form and behavior of instance sets with similar properties. The identifier and value of the instance meet the descriptor, the examples of the model appear primarily in line with the descriptor model, identification (Identity), Link, object (Object). Semantic instance with an identifier.
In other words, at different times of the system run, even if the value of the instance changes, the instance can be identified by the same example of different time points. At any time, the value of the instance can be a data value or a pointer to perform other instances. The data value is degraded, the identifier is the same as its value, and it can be considered from another angle. In addition to the identifiers and values, each instance comes with a descriptor, which is used to constrain the values that the instance may take. The descriptor is a model element for explaining instances. Most model elements in UML have such a double feature, and the main content of most elements is various descriptors. The purpose of this model is to illustrate the possible different values of the system with the values of the instances and instances of the system. Each descriptor describes an example. Objects are instances of classes; links are associated instances. Use an example to illustrate possible use case instances; parameters illustrate possible variable metadets, so push. Some instances do not have a family name unless it is formally set, it is often overlooked, but this does not exclude their existence. If one of the possible possible appearance of the state in which the state is executed, if each instance in the system is an instance of a description word in the model, and the instance set meets all explicit and implicit constraints in the model, the system value is legal. . Model Description System Possible Values and the Behavior of Systems from a value from a value to another during execution. The value of the system is a collection of all instances in the system and their respective values. The behavior element in the model shows how the instance therein is transitioned from a value to another value. The concept of an instance identifier is important to this description. Each behavior step illustrates a change in a few instance values in its previous value. The rest of the system is unchanged. For example, the local operation on the object can be represented by an expression of the new value for each attribute, while the rest of the system is unchanged. Non-locale functions can be broken down into partial functions on several objects. Note that the instance in the execution system is not a model element, usually, they are not even as part of the model. The example in the model is a short capture of a typical structure and behavior. It is a short capture of the system to validate or run historical trajectory, which is useful for people's understanding, but they usually do not define anything, just indicate many possible Value. Direct instances each object is a direct instance of a class and an indirect instance of the ancestors of this class. The same is true for other optically generalized elements, if a class descriptions an object and this object is no longer illustrated, the object is a direct instance of such a class. In multiple categories, one example can be a direct instance of multiple types, and there is no ancestral relationship between these types. Some execution semantics, by a class element indicating the class, other indicated roles or types. The full descriptor describes all attributes, operations, associations, and other features of the instance, which can be obtained from the direct class or from the ancestors. When multiple categroups, the full descriptor is a combination of the properties of each direct class. Creating an instantiation described in an instance creation method. Although the descriptor is different from the example, they have many common features including the same format (because the descriptor must explain the format of the instance). Therefore, the corresponding relationship should be reflected as much as possible for the descriptor - instance to select the representation of the selection. There are not many ways to achieve, each has advantages and disadvantages. In UML, the two use the same geometric symbols, and the name characters of the instance elements are joined. Although FIGS. 13-110 represent objects, the underscore representation is applicable to other instances, such as examples of the use example, component instance, and node instances. Since the example in the model is as an example, it is usually only included with the details related to specific examples. For example, you don't have to list a complete property table; if you pay attention to other things (such as information flow among the information), you can even ignore the entire property table. FIG. 110 describes the relationship between an instance of a descriptor and an instance of a descriptor. See example (Instance). 173. Instantiable can have an example, synonym: Concrete.
See abstract, direct instances, generalizable element. Semantic generalization elements can be declared as abstract or can be instantiated. If they are instantiated, you can create a direct instance. 174. Instantian Create an instance of a descriptor. See instantiation. 175. Instantiation's creation of new model elements instances. See initialization. Semantic instances are created by original creation activities or creation operations at runtime. First generate an identifier for a new instance, then generate a data structure according to the description of its descriptor, and finally the initialization of its attribute value by the descriptor and the creation operator. Instantian use dependencies represent the relationship between classes that create instances or include this operations and instantiated objects belonging. Object When an object is instantiated (created), the identifier and the storage area are given, and the object is initialized. The initialization of the object defines its attribute value, association, and control. Typically, a specific class has one or more types of constructive operations to create a new object of such a class, and all constructors potentially create a new raw instance, which is then initialized by the constructor. After the raw instance is generated, it has the format specified in the descriptor, but its value is not initialized. Therefore, the instance cannot be used by the system before initialization. Links Similarly, links are created by creating activities or operations. It is usually a scope operation of the corresponding class, not the constructor of the associated element itself (although this is also a possible implementation technology). There is also a potential operation of a new link between an object. If you have a similar association between a set of objects, the operation is invalid (because the contents of the associated content are a collection that cannot have iterative values). There is no further operation of the normal association, but links to associated classes must also initialize their properties. Example instance use case instantiation means that create an example instance, and use an example instance to begin execution. Example example may temporarily perform other use cases with extended relationships or include relationships, then return to the original use case. When executed to the end of the corresponding example, the use case instance will end. Other examples of other descriptors can also be generated by similar two steps: first create a raw instance, establish an identifier and assign data structures; pressing the value of the new instance to satisfy the corresponding constraint. For example, the activity is the direct result of the call operation. The authentication of the entity is the task of the runtime environment. The representation of the representation of the representation represents a dotted arrow, pointing from the actualized class or operation to the instantiated class; the arrow indicates that the structure type "instantiate" discussion sometimes uses instantiation to bind the template to generate binding element. But the binding relationship is more suitable for this. 176. The formal description of the intent descriptor structure and behavioral characteristics. Contrast: epitaxial. See the descriptor. Semantic descriptors (such as classes and relationships) are existing (connotation), and there are examples thereof (extension). The purpose of the connotation is to illustrate the structure and behavior of the instance with the manner that can be performed. 177. Interaction indicates how objects or other instances pass messages, and interactively define in the context of cooperation. See interactions (INRACTION DIAGRAM). Objects or other instances in semantic cooperation communicate through information exchange in order to complete a task (such as performing an operation). Information can include signals, calls. In order to complete a specific purpose, the mode of information exchange is called interaction. Structural interaction is a series of information exchange between a set of information (such as completing an operation). To illustrate interactions, you must first explain cooperation - that is, the object and its relationship is defined. The possible interaction sequence will be described later.
You can use a single sketch or more description with conditions (branch or conditional signals) (each descriptor instructions for explaining one of the possible execution paths), the complete description of the cooperative behavior can be implemented with a state machine, and its state is operation or other The execution status of the process. The representation of the method interaction or cooperation is represented, and both diagrams can represent the implementation of cooperation. The sequence diagram explicitly explicitly represents the behavior of cooperation, including the time order of information, and explicit representation of method storm activities. However, the sequence diagram only indicates the participating objects and does not represent their relationship and attributes of other objects. Therefore, it cannot fully represent the contextual view of cooperation. The cooperation diagram represents the integrity of the interaction, including objects and their relationships, so it is more applicable to the design. 178. Interaction Diagram is applied to a variety of views that focus on object interaction. Including cooperation graphs and sequence diagrams, both are closely related to the activity diagram. See COLLABORATION, Intertion. The method indicating the interaction between the objects in the method. The interaction map develops to different forms on the basis of the same information, each has different side focus. They are: sequential diagrams, cooperation drawings, and event maps. The sequence diagram represents interaction with time, focusing on the live lines of the interactive objects and their exchange of information. The sequence diagram does not represent links between objects. Depending on the purpose, the sequence diagram has different forms. The sequence diagram has a general form (describing the possible sequences), as well as examples (describes one execution sequence consistent with the general form). In the case where there is no branch and cycle, both forms are assigneous, and the descriptor is the prototype of the same example. The cooperation diagram represents the interaction between the objects that perform the operation. It is similar to the object map, indicating that the objects required to implement high-level operations and the links thereof. The time order of information is represented by the sequence number on the information flow arrow. Synchronous and asynchronous signals can be represented by a corresponding syntax. The geometric order of the arrow is represented in the geometric order of the sequence diagram, so it is not available. Sometimes the serial number is used in the sequence diagram for convenience, or to allow switching to the cooperation diagram. The sequence diagram and cooperation diagram show a similar information in different ways. The sequence diagram represents the exact order of the information, which is more suitable for real-time descriptions and complex situations. The cooperation diagram represents the relationship between the objects; it is more suitable for all the functions and process designs of the object. The discussion activity diagram represents the process steps in performing high-level operations. It is not interactive, indicating that the control flow between the process step is not the control flow between the objects. The side focus is the step in the process. It does not represent the task of the operation of the target class. The active map constructs a state machine that represents the state. Many special icons in the activity chart are equivalent to UML basic structures, just limited to additional limits, providing a well-species representation. 179. Interaction View indicates a view of information exchange between the objects to complete a task, including cooperation and interaction, with a cooperation diagram, and interactive. 180. Interface Description Elements The naming operation set of elemental behavior. See Classifier, Realization. The semantic interface is an externally visible operation of the class, component, or other entity (including the package and other summary unit) of the internal structure. Each interface only describes a limited portion of the actual behavior. A class can support multiple interfaces, effects or mutually exclusive, or overwritten. The interface is not implemented, lacking attributes, state, and associations; it only has received signals and operations, the interface can have an extension relationship, the sub-interface has all the ancestors of their ancestors and the received signal. The interface can have an extensive relationship. The sub-interface includes all of its parent operations and signals, but there can be additional operations. Interfaces and no properties, the method is equivalent to abstract operations. All operations in the interface are public visible (otherwise it is impossible to reference them, because the interface does not have a so-called "internal" to reference them). The following extended definitions. The interface is an operational set for explaining some of the service of the class or component. The interface is used to name a set of operations and indicate its signal and utility, the interface focuses on the service, not the structure.
The interface does not provide an implementation of the operation, and the operation list of the interface can also include a class that can be processed. The interface is used to illustrate the service provided by the server for other model elements. The interface is naming a set of operations, this set of exercises together to achieve part of the system or part of the system. The interface defines the service provided by the class or component, which is implemented by the class or component. Therefore, the interface has across the system's logic and physical boundaries. One or several classes (may be part of a component subsystem) can provide a logical implementation of an interface. One or several components can provide a physical package that conforms to the same interface. If the class implements an interface, it must declare or inherit all the operations of the interface, which can also have additional operations (see Implementation). If a class implements multiple interfaces, it must contain all the interfaces. Multiple interfaces can have the same operation. If the flag is in line with, it must be the same, otherwise the conflict will illustrate the system as non-rootic structure (implementation can be used to match the logo. For example, the parameter name and return type are ignored.). The interface does not declares the properties or relationship of the class; it is part of the implementation of the class. . The interface is an optically modified element, and the sub-interface inherits all the operations of the ancestors and can have new operations. Implementation is considered behavioral inheritance; the class inherits the operation of other types of elements, not the structure. A class can implement another class. It is like an interface that is like an interface, and only the operation part affects the relationship. Interface Participate. The interface cannot be used as the associated starting point, but can be associated with the target; as long as the association is only navigating to the interface. Indicates that the method is a type element, which can be represented by a rectangle with keyword "interface". The operations supported by the interface are listed in the operational tab. There is also a signal in the operating point (with the structure type "Signal"). Signals can also be listed in stand-alone. Attribute points can be omitted because it is always empty. The interface can also be expressed as a small circle, the interface name is below the circle. The circle symbol is connected to the class or other element of the support interface with a solid line, and it can also connect to a high-level unit, such as a package containing classes. This shows all the operations of the class support interface type (possibly more). The circle representation does not represent the operation of the interface support, and the operation list is indicated by a complete rectangular representation. The dotted line connects the interface and the class that uses its operation, and the arrow points to the circle. The dotted arrow indicates the operation declared in the class, but the user class does not require all the operations in the interface. Service usually uses the effect test to explain that if the provider provides an operation in an interface, the customer needs to meet the needs of the customer. Implementation Relationship with a dotted line indicating a solid triangle arrow (a broken line generalization relationship mark), the arrow points from the class to the interface it supported. This is the same as that indicates the type of type to achieve the class. In fact, this symbol can be used between any two types, indicating all operations defined by the customer (arrow) support provider (arrow), but does not have to support the server's data structure (attributes and association) Examples Figure 13 111 is a simplified view of the business components of the fiscal financial position. FINANCIALPLANNER is a personal finance application that records personal spending and investment. It has the ability to refresh the price of securities. MutualFundanalyzer carefully examines the public fund and needs to refresh the current price of securities prices. The ability to refresh the price is expressed with interface Updateprice. There are two components to implement this interface, indicating them to the solid lines connected to the interface symbol. The component ManualPricEentry allows the user to manually enter the price of the selected securities. QuoteQuery gives the price of securities from the quotation server by modem or network. Figure 13-111 The provider of the interface and the client of Figure 13-112 represents a complete form of an interface as a keyword of the class. This interface has two operations - ask the price of securities and obtains a value, submit a price list for securities and acceptance, in this case, QuoteQuery is connected to the interface with the real line arrow, this with the previous picture The same relationship is the same, but it is just more exact.
There is also a new interface PeriodicUpdatePrices, which is the original interface. It inherits two original operations and adds an operation and automatically refreshes the price. This interface is implemented by the component quoteserver. It also implements two operations in QuoteQuery, but the method is different. It does not share the implementation of QuoteQuery (in this example), so it does not inherit its reality. Figure 13-112 shows the difference between interface inheritance and complete inheritance. The latter includes the former, but the sub-interface can be implemented with a different method with the parent interface. QuoteServer supports the interface (ie UPDATEPRICES) implemented by QuoteQuery, but does not inherit the QuoteQueryde implementation (usually, while inheriting implementation and interface more convenient, so two structures Often the equivalent) interface can include a list of signals it can process (Figures 13-113). Figure 13-112 The full interface indicates that the interface interface with signals with signals in Figures 13-113 is used to define behaviors of classes, components, without limiting their implementation. In this way, the interface inheritance is allowed to separate, such as the declaration of the Java language. 181. Interface Inheritance Inherits the father's interface rather than its implementation or data structure. Vision for supporting interfaces without inherit implementation is implemented by using implementation relationships. Note that in UML, the generalization relationship implies an interface and implementation inheritance. Only by private inheritance can guarantee only only inherited examples without inheriting the interface. See Implement Inheritance, inheritance, private inheritance, realization. 182. Interface Description For an explanation of the behavior of the associated class, this behavior is used to meet the associated content. It includes a pointer to the interface, class, or other type of element that illustrates the required behavior. See the association role, a RoleName, a type (TYPE). Semantic usually, the ability of associated classes not only supports an association. For example, classes may participate in other associations, and their behavior is the sum of the behaviors required for all associations. A class should be more clearly described in a class that supports an associated function. The interface specifier is a class element on the associated end for explaining the functions used to support this association without considering the functions for other associated uses in the target class. Interface specifiers are not required, in most cases, a class is only participated in an association, there is no excess function. If the interface specifier is omitted, the association has access to all functions of the associated class. The specifier can be a set of types, each of which illustrates the operations that the target class must support. The target class must satisfy all requirements, it is possible to be higher than the description requirements. Representation of the representation of the syntax character name: iname. INAME is the name of an interface or other type of element instead of a role name. If there are multiple explanatory classes, their names are separated by commas. Example: In Figure 13-114, the Server layer stores the request in an Array class. For this, it only requires the Queue class function, not random access. The actual Array class meets the requirements of the interface specifier Queue (including an array of queue functions), and Monitor uses an array to represent the status of the request, using all the features of Array. Figure 13-114 Interface Description Discussion Using Role Names and Interface Description Is To create a cooperation, where only one associated role and two kinds of element characters are included, and their structure is defined by the original role name and role class element, so cooperation uses associations and classes. The original class must comply with the interface description (which can be interface and type) 183. Internal Transition status has an active conversion, but it does not cause a status change. See the state machine (State Machine).
Semantic internal conversion allows events to cause activities while do not change the state, it has initial state without the target state, once activated, the activity on the conversion will be executed, but the status will not change. Therefore, no entry or export activity is performed. From this regard, it is different from self-conversion, and self-conversion will cause exit nested state and trigger an export or inlet action. The indication of the internal conversion text entries is the same in the internal conversion division of the state, and the entry is the same as the text tag semantics of the external conversion. Because there is no target state, you don't need to use an arrow. Event Name / Active Expression Entry, Exit, Do and include is a reserved word, which cannot be used as an event name, which is used to declare entry activity, export activity, activity execution or sub-mechanism execution. These special activities use internal conversion syntax for more obvious. They are not internal conversion, which is not an event name. Figure 13-11 shows the representation. Discussion Internal conversion can be considered a "interrupt" that triggers an activity but does not change the current state, so no export or inlet action is activated. The best way is to model an activity by connecting an internal conversion to a constituent state, the activity must appear in some states, but may not change the status of the status - such as means of helping information or events. It is not advisable to model the modeling or abnormality, and they should be modeled with transitions to new states because their appearance does not match the current state. 184. INVARIANT a must always constrain true constraint (or at least true at all activities). Semantic invariance is a true Boolean expression that must be true in any time without active operation. It is a statement that statements instead of executable, depending on the form of expression, it can or cannot be determined in advance. See the precondition, a postcondition structure constant as a constraint, with a structure type "invariant". Indicates that the method of subsequent conditions can be expressed as a note with the "Invariant" keyword. Note Attached to class, attributes, or other elements. 185. Iteration Expression generates an expression of an iterative case. Each iterative use case specifies an execution of one activity in iteration. Iterative cases can include assignment to iterative variables. Each iterative case is only executed once. See the message (Message) sessile representative representative or iteration execution. According to the conditions included, 0 or more messages are shown. Select: * [Iterative Sentence] A Iterative [Condition Sector] A branch an iteration represents a message sequence, iterative clause represents iterative variables and test details, but can be omitted (in the case where no iterative conditions). Iterative subseerators express language to design language with pseudo code or actual programming. UML did not explain its format. Can be * [i: = 1 ... n] The execution of the message represented by the condition is depends on whether the conditional clause is true. Conditional subseevice use of pseudo code or programming language expression. UML did not explain its format, which can be [X> Y], and the branch is the same as the iterative representation without an asterisk, and it can be considered only once an iteration. It is possible to execute in the messages in the iteration, which may also be executed in parallel, and its representation is an asterisk plus double vertical line. For example: * [i: = 1 ... n] || Q [i] .calculatescore () Note that in the nested control structure, the representative decree is not discrete in the internal level of the serial number, each layer defines itself in its context. Iteration. 186. Iterative Development (Item Development) includes a series of steps for system development, or iterations. Each iteration is more close to the desired system than the previous. Each step must be an executable system for execution, testing, and debugging. Iterative development is in conforming to progress development. In an iterative progress development, each iteration adds new capabilities to the previous. The order of capacity increases should take into account iterative balance and find significant risks as soon as possible. See the Development Process. 187. Binding (JOIN) state machine activity map or sequence of positions, there are two or more side-by-side threads or states to be attributed to a thread or state; a binding or non-branch.
Antonym: Branch See complex transformation, composite status. Semantic combination is a plurality of source states and a transformation of a target state. Once all source states are active and trigger events, conversion, the target status is active, and the source status must be in different regions of the parallel composition state. The representation is expressed as a plurality of transition arrows and a thick line that rotates the arrow. It can have a conversion tag (monitoring condition, trigger event, and activity), Figure 13-116 is an explicit bond from multiple states in parallel composition states. Discuss see the merger. 188. Junction State state is a integrated conversion of a partial pseudo state. It does not interrupt running to the completion step in the conversion execution. See branch, merge (MERGE). The conversion in the semantic state machine can reach the target state from the boundary of several components. One or more export / entrance activities are activated when performing this conversion. In some cases, it is necessary to convert the export / entrance activity of one or more activities and nested states. This is impossible for only a simple conversion of only a single activity. There will be a single output or allowing a number of outputs that trigger a number of guardians that will be triggered. The combined state is a pseudo-state such that a comprehensive conversion is established from multiple conversion segments. The binding state can have one or more transfer segments and one or more transfer segments. It does not have internal activities, subconstruction, or outward conversions with trigger events. It is a pseudo-state for building a converted, so that it cannot be in the "activity" state. The binding state can be constructed from multiple pieces of conversion. Only the first segment in the binding status chain can have a trigger event, but each segment can have a guard condition. There must be no contact with the segment. Effective monitoring conditions are a combination of all single guard conditions. The conversion will not start unless all conditions are met. In other words, the state machine is not in the binding state. If multiple conversion enters a combination, they each can have different triggers, or no trigger. A different conversion is representative by each path of the combined state set. Outover conversion can also have a care condition. If there are multiple outward conversions, each must have a single guard condition, which is a branch. Outover conversion can come with a activity (the combination state can have an internal activity, but it is equivalent to connecting an activity to an outwardly). Once all guardings are met, the activity will be executed. The conversion cannot be partially excited and stopped in the binding state, and a regular state must be reached. When the transfer is activated, the outward conversion will be immediately activated, and the included activity is performed. Transition to and outward is a portion of an atom step (running to the end step) - it cannot be interrupted by any activity or event. The binding state is indicated by a small circle in the state machine. It has no name, which can have an inward and outward conversion arrow. Example: Figures 13-117 are two complete conversion of state S to T - a single conversion and event E triggering multiple pieces of event E triggering. Inlet / export activity is also shown in the figure. Note that the location of the activity label on the conversion line is meaningless. If the active D draws in status X, it will execute before x exit Y. Therefore, D should be painted at the outermost position on the conversion. Figures 13-179 and Figure 13-184 show other examples. See the status map of the control icon and the symbol of the symbols that may appear in the activity diagram 189. Keyword keyword is a text for model elements that do not have independent grammar. See Graphic Marker, a STEREOTYPE indicator keyword is used to have a built-in model element and a user-defined constructor of a separate representation. The keyword is usually written within the book name. "Keyword" If the symbol with the keyword has a certain size, hit the keyword in the symbol boundary. Some predefined keywords are described herein and is represented by reserved words. Others can be used as constructive types. The same structure type as the predefined keyword cannot be used. The symbols that are easy to distinguish are limited, so the UML represents the text keyword to distinguish the different variations under the same topics, including the meta-class type subclass, meta-model base type structure type, and table element group.
For the user, the elementary model between the metamorphic subclass and constructive type is not important, but it is important for tool manufacturers and realizing meta models. 190. Label uses a method of using a string in the figure, is merely a representation. See the picture (Diagram). The representation tag is a graphic string that is logically attached to another symbol in the figure. This attachment is usually indicated by the character string on the side of the symbol or the closed area. For some symbols, the label position is determined (such as online). For most symbols, the label must be close to the line or target. The programming tool maintains internal contact between the label and graphics symbols, so that both the label and graphic symbols are indicated separately, both are logically connected. It must be clear in the final view to ensure that symbols and tag connection relationships will not be confused. Although the map is not fully clearly clearly clear, this relationship is clearly no objection in the structural level of the figure. The tool can use other methods to represent the accessory relationship between the tag and the symbol (such as a coloring line or matching element flicker). 191. Language Type uses an anonymous data type defined by programming language. See the data type (Data type). Semantic types are an expression that is translated into data types by programming languages. It can be attribute, variable, and parameter type. It does not have a name, which cannot be used to declare the new data type. For example, the C data type "Person * (*)" can be defined as a C language type. The purpose of the language type is the implementation of the programming language. More logical relationships should be implemented with association. 192. Layer A structural mode of the same abstraction hierarchy in the model, each layer represents a virtual world at a level of implementation. 193. Leaf (Leaf) In the generalized hierarchy, there is no lypel element, which must be a specific (overall implementation) entity for use. See abstraction (concrete). Semantic leaf attributes declares that an element must be leaf. If the child is declared, the model is a non-rootic structure. Its purpose is to ensure that classes are not modified. For example, the behavior in the class must be complete, so there is practical meaning. The leaf declaration also allows the compilation of the system to different parts, which guarantees that the method code is not overloaded, and the embedding of the convenience code. The element of the leaf property may be actually leaf, but it can generate a child when modifying the model in the future. It is not a basic semantic property that is considered to be a leaf or a leaf is not a basic semantic property. 194. The dotted line in the lifeline sequence diagram, is used to indicate that the object exists within a period of time, and this line is parallel to the time axis. See sepuence diagram. The semantic lifeline represents the time period of the object. Objects are activated when there is a control thread - that is, as the root of the thread. Passive objects temporarily exist when the control thread is passed - is called outside. The latter is called activity, and its existence time includes the time of the process call the lower process. The representation object or class is expressed as a vertical dashed line in the sequence diagram, referred to as a lifeline. Life line indicates that the object exists within a specific period of time. The arrow between the lifeline represents the message between the object, pointing to the lifeline arrow to represent the object reception information, which is completed, and the arrow indicates the object transmission information, activated by one operation. The geometry of the arrow between life lines represents the time order of the message. If the object is created or died within the time period of the sequence diagram, its lifeline starts or terminates at the corresponding time point. Otherwise, the lifeline penetration is always. The object symbol is drawn on the lifeline. If the object is created in the figure, the object symbol draws on the message created, otherwise, the object symbol is drawn on any message arrow. If the object is destroyed in the figure, use large x to indicate destruction, X or marked on an arrow that causes the destroyed message, or the final return message of the destroyed object (self-destruction). The object exists at the beginning of the conversion begins on the top of the figure (above the first arrow). The object still exists after the conversion, the lifeline draws the last arrow. Life lines can be divided into two or more to indicate conditional control.
Each trajectory corresponds to a branch of the message flow. Life lines can also be combined at a certain point, see Figure 13-162. This representation is to be confused and should be careful. The life line is represented by a solid line permanently or temporarily active. The second two-line represents iteration. Xiang you will activate. Because the active object is always active, sometimes the double line is saved. Life line is interrupted by state symbols to indicate status conversion, which is corresponding to the conversion in the cooperation map. The arrow that points to the state indicates a message that causes a state change. See Figure 13-163195. Connection (LINK) Object Reference Unit, is an example of associated or associated roles. Semantic connections are independent connections between two or more objects, which is an object reference group (ordered table), which is an instance of association. Objects must be a direct or indirect instance of the corresponding location in the association. An association cannot have an iterative connection from the same associated, that is, two identical object reference tuples. As an associated class instance, there is an attribute value table in addition to the object reference element group. It is not allowed to have an iterative connection with the same tuple object reference, regardless of whether the attribute is different. The identifier of the connection is from the object reference tuple, which must be unique. The connection can be used to navigate. In other words, the object connected to the other end can get the object of the other, and the message can be sent (referred to by contacting the send message). This process is effective if the connection has navigation in the target direction. If the connection is not navigable, access may be valid or invalid, but the message transmission is usually invalid, and the navigation in the opposite direction is additionally defined. In cooperation, the associated role is the relationship between the temporary type dice related to the context, and the example of the relationship role is also connected, and its life is limited to the length of cooperation. The representation is expressed as the path between the object - one or more connected lines or arcs. In variable correlation, the path is a loop pointing to the same object at both ends. Represents the name of the instance to underline. The connection has no instance name, its identifier comes from the related object, because the instance does not have multiple sex, and there is no multiple. Multipleness is an identifier that describes the number of instances that can be present. Other associated modifications (aggregation, composition, navigation) can be represented on the connection role. Connection can represent qualifiers, the value of the defined word can be represented in the box, and Figures 13-118 have a conventional connection and a qualified word connection. The connection can also represent the associated attribute, such as the navigation direction, aggregation, composition, and implement the structure type and can be implemented. Vision. N One-dimensionally connected N-dimensional connection with diamonds, which leads to the path to each participating object, and other representations are the same as above. How to discuss how the object map represents a dependency? Typically, dependent on the relationship between the class and is represented in the class diagram, not in the object map. Process variables, local variables, and operation calls should appear as actual data types and not just dependence. Therefore, they can represent a pointer to connect the caller to get the target object process - this is the connection. Some connections are examples of associated roles in cooperation, such as variables and local variables. Remember: Relying on the class, not related to each object. 196. An instance of the connection end associated with the connection end. 197. List (List) is a sequence of ordered an ordered model element sequence that has a variable length of another model element. See Classifier, State. Semantic class has many sub-lists, including properties, operations, and methods. There is an internal conversion list. Other elements have other elements list. Each list is explained alone. This technique is usually used to explain the properties of the table. In addition to the attributes and operations lists, other list of other lists are available, such as responsibility, role, or modification history. UML does not define these optional lists, depending on the user-defined list operation depends on the tool. Embed list and the elements in which you are included, or you belong to other contained elements. Multiple contained elements cannot share ownership. Other classes can access elements in the list -, for example, by inheritance or association, but only directly containing elements have all rights and modifications. These list elements are stored, copied, and destroyed together with elements. The order of the list elements is determined by the modeling. This order will be useful - for example, as a code generator that generates a programming language declaration sequence.
If the moderator does not care, it may be, because the model is in the analysis phase or because the language is independent of the sequence, this order is still existed but it can be ignored. The representation is embedded in a list of independent points as a string list, each element corresponding to a string. Each string is a code representation, such as attribute, operation, internal conversion, and more. The coding type is described by each element. The order of the sort string and the order in the list of elements in the list, but the internal storage order may not be used, and the string storage may depend on some internal properties, such as name, visibility, or constructive type. Note: Each is still maintained in the order in the basic model. Sorting information is only ignored in the view. The omitting number (...) is the element representation of the element indicating that the model is also qualified elements, but not listed in the table. These elements may be listed in other views. Construction type list elements can have a structure type. Texture Type Keyword Written in front of the list element. Property string Property string Describes the properties of the element. In the braces after the elements, the properties and constraints are listed with commas. Group properties A list of list elements can have attributes or constructive types. This line does not represent a list element if the constructed type, keyword, attribute string, or constraint appears in a separate line. Instead, all the elements are common, as they act directly on each line. The default effect continues until another group of properties in the appearance table. Insert a row with an empty keyword ("") to undo all group properties, but will not be limited to all items of the group properties, will be more clear. Figures 13-119 show a configuration type acting on a plurality of elements. Note: Group properties are only a convenient representation, and each model element still has its own attribute value. The fractal name points can have names, indicated by special fonts (such as lowercase black bodies), labeled them on the top. This flag is very useful if there is an omitted or new user-defined point. For classes, predefined points are called Attributes and Operations. User-defined points may be requiremements. The name of the class must be, so the tandem name is not allowed. Figures 13-119 and Figure 13-120 show a naming point. Indicates that the option sorting tool may list the table elements by the storage order. At this time, the inheritance sequence of the element can not be seen. The order is based on some internal properties, and no additional modeling information is indicated. Typical sorting includes alphabetical order, constructive type sequence (constructor, destructuring function, and general method), sorted by visualizable (public, protected to private), and so on. Filter list elements can be filtered in accordance with certain rules. The instructions for selecting rules are the responsibility of the tool. If the filter list is empty, it means that the elements do not meet the requirements, but the original table may or may not contain other elements that do not satisfy the rules, which are not visible, the tool should be responsible for explaining partial and all filtering and how to explain. This will also be described with separate graphs. If a point is omitted, the existence or not of the elements of the elements cannot be determined. Empty points indicate elements that do not match the screens. Note: Properties can also be generated (Figures 13-71) 198. Location, a runtime entity in a certain environment, such as objects or points in a distributed environment. In UML, the location is dispersed, and the location unit is a node. See Components, Node (NODE). The concept of semantic location requires the concept of the entity. UML does not establish a three-dimensional space, but provides a topological model space connected by a communication path. The node is the computing resource that can exist when the runtime entity. The node is connected to the communication path, and the position of the entity is the assignment node pointer. At nodes, some entities exist within other embedded entities. For example, an object exists in a component or other object, where the position of these entities is the position therebetween. Objects or component instances can be moved to a new location.
This can be expressed by "Become" relationship, "Be" means that the first entity is replaced by the second entity at a certain moment, while the second entity is originally in another position. The position indicating that the method (including the object, component entity, and node entity) can be used in another entity, as shown in Figures 13-121. The relationship is also expressed in constituent arrows. Alternatively, an instance may have an attribute of a Location tag whose value is a name containing an element. If the object is moved in interaction, it can be represented as multiple versions, and the version is connected to the Become. See Figure 13-121. The Become arrow can have a serial number, indicating the time of the object movement. Each object symbol represents a version of the time period object. The message must be connected to the correct object version (Figure 13-117). 199. Many (MANY) Multipleness 0.. Abbreviation --0 or unlimited multiple. In other words, the size is not limited. See MultiPlicity 200. Members 'names' names of the named structure inheritance organization, which can be attributes, operations, or methods. Each class can have 0 or more members. A list of specific types of members represents in the form of a class element symbol in the form of a string table. See the list (List). 201. Merge (MERGE) state machine, two or more alternative control paths are in this converge or "branchless". Antonym: branch. See the Junction State semantic merge to refer to the case of two or more control path convergence. In the state machine, one state has multiple transfer conversions. There is no constructed specific model constructed. If it is a single run to a part of the completion step, it can be represented by a combination state. The representation is in the state diagram, sequential diagram, or active diagram. The combination is expressed as a diamond with multiple input arrows and an output arrow. No conditions are required, see Figure 13-122. Rhooles are also used for branches (contrary to the merger), but will not be confused. Because the branch has an input arrow and multiple output arrows and has a guard condition. It can be combined with branches and merge, but it is not used. It can have multiple inputs and multiple tagged outputs. Note: Merger is just a convenient representation, omitting it will not lose information. Merge and branches are often used. Figure 13-122 Combined discussion Please distinguish between merger and binding, combined with more than two control paths. In any execution, just take one each time, no need to synchronize. A parallel control path is combined with two or more. In any execution, all paths have to be passed, when they reach the binding source state, the combination is activated. 202. Message (Message) passes the message from an object (or other instance) to another, unless the activity can guarantee this. The message can be a signal or call operation, and the message instance is considered an instance of receiving events. See call (COLLABORATION), Interaction, Operation, Send, Signal. The semantic message is to send a signal from an object (sender) to another or several other objects (recipients), or call another object (sender or caller) to call another object (recipient). The implementation of the message has different ways, such as process calls, internal communication between the active thread, and the occurrence of events. At the logic level, the sending signal and calling operation is similar, all activates communication between the sender and the recipient, and the recipient receives a value and thus determines what should be done. The call can be considered as a signal that activates a transmission, the explicit belt has a pointer change, and the signal returned back to the return value by it. At the level, the signals and calls have different attributes and behaviors, so they are different UML elements. The received signal may cause the conversion of the recipient state machine.
The recipient has two different call processing methods to be selected (determined by the recipient model). The operation is implemented as a process body (method) that it will be activated when the signal arrives. After the process is executed, the caller recovered the control and the return value can be retracted. Another way is an active object, and an operation call may cause a calling event that triggers a state machine conversion. In this case, there is no method body, but the conversion can have actions, and the conversion can also provide a returning value for the caller. When the conversion is completed, the caller reclaimed control when the event is not triggered. A collection of target object expressions in the message. Message all objects in the collection. Unless otherwise stated (constrained), the message will be sent in parallel to all objects. This shows that the order in which the execution is unpacked, and may be parallel. If you must send messages in a certain order, they should be looped. For calls, all calls are completed, the caller reclaimed control. The time to send messages can be represented by an expression on the message name. See the Timing Mark. Structural messages include senders, recipients, and activities. In interaction, the sender is a class element role that issues a message. The recipient is a class element role that receives the message. The activity is called, signal, partial operations or original activities of the sender, such as creation or destruction. The event has a parameter table, a list of recipients, and a pointer to the activated operation or signal. You can also have a message or iteration description. In interaction, there is a pre-drive-subsequent relationship and call-called relationship, which is applicable to process methods. Add a nested level each time you call. In the call, the message is ordered, it can be parallel. Pre-drive-Subsequent (order) relationship will be a column on the message organization on the thread. One message can have multiple pre-drive or successive. If there is no order relationship between the two messages and there is no common pre-drive, they can work in parallel. If a message has multiple front drives, the message can only be executed after all the pre-drive is complete. This message is a synchronization point. Call - The call (activation) relationship defines the nesting process structure. The message of the calling process (using the calling event) is the activator of all messages in the process of the caller. Wherein, the called message also has a pre-drive-subsequent relationship, thereby determining the relative order (allowing parallel). If the message is called, the caller will lock before the calling process is completed and returned. If the recipient processes the message to call the event, the return value will appear after the initial conversion is complete, and then the caller reclaimed control and continued to execute by the calling process. The order and activation relationship only involve the message in the same interaction. Indicates the representation of the sequence diagram and cooperation map. In the sequence diagram sequence diagram, the message is represented as a solid line arrow from the lifeline of another object (sender) from an object (sender). Compared to external messages, if the arrows and lifelines are vertical, the message transmission is immediate, at least very fast. If the arrow is tilted, the transmission of the message has a certain time delay, and other messages can be transmitted therebetween. The object sent to itself, the starting point and end point of the arrow are the same lifeline. The message is arranged vertically from the top to the end in chronological order. If multiple messages are parallel, the order between them is not important. The message can be serial number, but because the order is represented by a relative relationship, the serial number is usually omitted. Passing the delay message arrow is usually horizontal, indicating that the delivery message is short, and there is no "other events during this period. For most calculations, this is the correct hypothesis. If the transmission of the message takes a certain period of time, other events (from the other party arrive) can occur during this period, the message arrow can be sloped downward. The branch issues multiple arrows from the same point to represent branches, each with a guard condition. Whether the conditions are mutually exclusive, there are two structures and parallel types.
A series of messages that are repeated can be encapsulated and labeled repeated, and repeated flags indicate that this group message can appear multiple times. For the process, repeated conditions are listed below the repetition. If there is a parallel, some messages may have a repeated part, some only execute it. Collaboration Diagrams In the cooperation map, the message is indicated as an arrow with the label, attached to the path to the sender and the recipient. The path is used to access the target object, and the arrow points along the path to the recipient. If the message is sent to the object itself, the arrow points to the object yourself, and is marked with the keyword "self". A plurality of messages can be connected to pass along the same or different paths. The order is indicated by the serial number. Double View (Both Diagrams "message arrow labeled a message name (operation or signal name) and parameter values. There is also a serial number on the message, indicating the role of their interaction. The physical location of the arrow in the sequence diagram indicates that the sequence number can be omitted. There must be serial number in the cooperation map. The serial number can indicate the parallel thread, which is useful for both images. The news can also have a monitoring condition. Control flow Type The following different arrow shapes represent different control flow. Solid solid line arrow process call or other nested control. All inner sequences should be completed before the outer sequence is from the new beginning. Can be used for conventional process calls. It can also be used in parallel activity. Indicates that one of the states transmits the signal and is completed in the sequence of the nested behavior. Stop arrow plane control flow, each arrow indicates the next execution order. For nested processes, it corresponds to the bottom scan activity of the search leaf node. Half stick arrow asynchronously control flow. Unlike the stick arrow, it is used to indicate asynchronous messages in order between two objects. The dotted stick arrow returns from the process call, and the return arrow can be ignored. Because it is implied at the end of the event. Other variants can represent other types of control, such as "timeout", "abort", etc. These are extensions outside the UML core. The message tag syntax is as follows: Tag Description Send message, parameters, and return values, as well as the order in large interactions, including call nested, branch, branch, side, and synchronization, etc. In the pre-drive cooperation, the serial number listed in the pre-drive was separated by a comma and then following the backlash (/). The serial number list / if the list is empty, then the entry is omitted. Each serial number is a sequential expression without any iteration, which must match the serial number of other messages. This means that the message can be sent after the message representative of all the serial numbers in the list (a message stream that is required, and the conditioning conditions are still satisfied). Therefore, the monitoring conditions represent thread synchronization. Note: The message of the serial number before a message is its default forward drive, which is not particularly listed. The same prefix message number constitutes a sequence. The digital pre-drive refers to the last bit of the serial number 1, that is, the message 3.1.4.5 is the pre-drive of 3.1.4.6. In the sequence diagram, the sequence is determined. The object sent itself before any messages sent itself, indicating synchronization. The serial number expression serial number expression is a list of sequence clauses, and between the colon (:) is separated. Each clause represents a nested level in interaction. If all control parallel, no nested. The syntax is as follows: The label loop OPT where the tag is an integer or the name integer represents the hierarchy. For example, messages 3.1.4 are after 3.1.3 of Activities 3.1. The name represents concurrent control threads. For example: messages 3.1a and 3.1b are parallel in 3.1. All parallel control threads are the same at the same nested level. Cycles represent conditions or iterations. Indicates 0 or more messages based on the conditions. Select: * [Iterative Sentence] Iterative [Condition Sentence] The branch iteration represents a message sequence on a given depth. [Iterative Sentence] can be omitted (not specified). [Conditional clause] is implemented in a dummy code or programming language. UML did not explain its form. Can be: * [i: = 1..n]. Conditions indicates whether the implementation of a message depends on whether the condition clause is established. Conditional subseevice implementation by pseudo code or programming language.
UML did not explain its form, for example: [x> y]. Note: The branch is the same as the iterative representation, just no *. It can be considered an iteration of the execution. The indicator assumes that the message is executed in order, or may be executed in parallel, indicate (* ||) with a double vertical line. Signature is a string of the name, attribute, operation return value, message or signal. Have the following attribute: Return Value list Description Interaction The name of the message returns the value, with a comma between the names. It can be used as a parameter of subsequent messages. If the message does not return a value, the return value and assignment operation are omitted. Message name target objects raised in the event name (usually an event requiring an operation). Can be implemented in different ways, one of which is the operation call. If the implementation is called, it is the operand. The operation must be defined or inherited in the class of the recipient. In other cases, it can be an event name that the recipient. Usually use the message name and the parameter table to determine one action. The parameter list brackets are used with a comma-separated parameter table. The blanket can also use parentheses. Each parameter is an expression written in a pseudo code or an appropriate programming language (not described in UML). Expressions can use the return value of the previous message (the same scope) and navigation expressions originating from the same object (ie, the link and the upward path thereof). Examples include Control Message Tag Syringe 2: Display (X, Y) Single Message 1.3.1: P: = Find (SPECS) Nested Call for Return Value [x <0] 4: Invert (X, Color) Condition Message 3.1 *: Update () iterations A3, B4 / C2: COPY (A, B) and other thread synchronization indication option indicates data tag on the side, which can replace the text expression of the parameters and the return value (Fig. 13-123). The label is a small circle marked with parameters or the return value, and the arrows on which the arrows are directed to the direction (parameter) or reverse direction (return value). The tag represents the return value and parameters. Both texts and tags can be used, but the text is more concise, it is recommended to use. Message syntax is described by programming languages, such as C or SmallTalk languages. But all expressions in the view should use the same syntax. 203. Metclass (Metaclass) This kind of example is class, usually used to construct a metamodel. The class can model the type of constructor with keyword "metaclass". See a power type (PowerType). 204. The meta-metamodel defines the model of the language used in the expression meta model. The relationship between the metammato and metamodes is similar to the relationship between the metamodes and the model. This level is usually related to tool builders, database builders. UML is a meta-Object-facility, referred to as MOF) element-meta-mode definition. 205. Metamodel defines the model of language used in the expression model, is an example of a metamm-meta model. The UML meta model defines the structure of the UML model. 206. The collective name of all entities in the metaobject element model language, such as the mid class, meta type, meta, meta, and metades. 207. MetareLationShip Relationship descriptor and the collective of its real case relationship, including instance relationships and strong types of relationships. 208. Implementation of method operations illustrate an algorithm or process of generating operation results. See the specific (operation), a realization semantic method is the implementation of the operation. If the operation is not abstract, it must have a method or call the event, or define on the class with an operation or inherit from the group. Methods The process expression is described, which is a string (such as C , SmallTalk, or natural language) for a particular language. The language should be adapted to the purpose. For example: natural language is suitable for early analysis, not applicable to generating code.
Operation declarations implies the presence of methods unless the operation is declared as abstract. In generalization inheritance, each iterative declaration is overloaded with any inheritance method in the same operation. Note that the method is an executable process - an algorithm - rather than a result description. For example, beforehand - afterwards, it is not a method. Methods are subject to algorithm, computational complexity and package implementation and address allocation issues. In some ways, the properties of the method are more stringent than their operation. The operation is not defined as a query, but the method can declare to query. If the operation is defined as a query, the method must be a query. Similarly, methods can strengthen parallel properties. The sequential operation can be implemented as a conditional control or parallel method. At this time, the method meets the statement of the operation, just enhances the constraints. The method is indicated by the operation statement, but there is no abstract feature (Fig. 13-124). If the operation is inherited, the general form of the presentation of the iterative operation (non-writing) is given, gives a specific operation. The text of the method is expressed as a comment, connects to the operation entry. The figures usually do not mean that the law is hidden, and the command line can be rendered. 209. The complete abstraction of the model of the model (Model) system. See the package (Subsystem). Semantic models are more or less complete abstraction from a particular viewpoint to the system. Say it is complete because it fully describes the system or entity from a given view. A model that provides a separate view can be maintained independently. The model can be broken down into hierarchy of the package. The outermost package corresponds to the entire system. The content of the model is a closed package containing (all) relationships from the top package to the model element. The model may also contain some of the system environment, which is expressed as a behavior and interface. Especially the correspondence between environment and system elements can be established. The system and its environment constitute a higher level. Elements in different models have not directly affected each other. But they may represent different levels, details, or development steps of the same concept. Therefore, the relationship between them, such as tracking or refinement, is very important to the development process, often affecting important design decisions. The representation method can be expressed as a package and has a structure type "model". However, there is no detail in the representation to represent the model. Tools can represent a list of models, but there are very few relationships. The most useful is to turn the model name into the name of its top package, or the map of all of them. Discussing no system view or system itself can be referred to as complete because the system is always associated with the outside world. Therefore, the package model is a relatively concept, which must specify the actual work needs. The UML model is a package that focuses on a view. Each model has its own level, which can be the same or different from the other view of the system. 210. Model Element Elements from modeling systems. Unlike expression elements, expression elements (usually visible) are one or several model elements interact with people. Semantic All semantic elements are model elements, including the concept of real world and concepts in computer systems. The graphic element for the purpose model is the expression element, not a model element because it does not contain semantic information. Model elements can have names, the use and limitations of the names of different types of elements, and will be explained below. Each model element has a named space that is consistent with its type. All model elements can have the following properties. Tag values Any model element or expression element can have 0 or more tag-value pairs. The label is the name of the meaning of the value. The label is not fixed in UML, but it can be extended to represent various modeling or programming tools. Constrained model elements can have 0 or more constraints. The constraint is a restriction condition that is expressed in a constrained language language. The structure type model element can be associated with 0 or more constructed type names, as long as the constructor is applied to the base model element. The structure type does not change the base class structure, but can increase the tag for an element with the type of construct, and the model element can participate in the dependencies.
See Chapter 14 Standard Elements, predefined tag tables (a list of predefined tags), constraints, strodypes. 211. This feature of the Model Management View model is a structured part of the structure-package, subsystem, and model. This view is sometimes used as part of a static view, which is usually used with a static view class. 212. Modeling Time appears in the modeling activity during software development, including analysis and design. Note: When discussing an object system, distinguishes modeling time and runtime are very important. See the development process, modeling steps (stages of model). 213. Module Software Unit, including source code modules, binary code modules, and executable code modules. It does not correspond to individual UML structures, but includes several structures. 214. Multiobject indicates a plurality of object collections rather than a single object of a single object. See Classifier Role, COLLABORATION, Message. Semantic multi-object is a class element role representing multiple objects, and multiple objects are typically associated with "many" terminals. Multi-objects are used in cooperation, indicating an operation of an object set rather than a single object. For example: Find an operation of an object in an object set to the entire collection rather than individual objects. This group does not affect the basic static model of the system. The representation of the multi-object is represented by two rectangles, and the above rectangles are biased downward, indicating a number of rectangles (Figs. 13-125). A message arrow pointing to a multi-object symbol indicates a message sent to these objects. For example: Find a selection operation of an object. To perform actions on each object in the related object, there are two messages: an iterative message generates a path to an object, and another message sends a message through these paths (temporary). The illustration can be simplified, the message combines to a one, including an iteration and the application of each object. The description of the target role name is "many" (*), indicating that there is a plurality of implicit connections. Although it can be written as a single message, it requires two layers (operation connection iteration, using connection messages) in the basic model (and in any actual code). Objects from the collections are represented as conventional object symbols and connected to multi-objects with a constituent chain, indicating that it is a collection. The arrow pointing to a single object symbol indicates a message sent to a single object. Typically, the selection message sent to multiple objects returns a pointer to a certain object, and then the original sender sends a message to it. 215 Multiple Classification generalized a semantic variant, one of which can directly belong to multiple classes. Semanticity It is a semantic variant, where the object can be a direct instance of multiple classes. When using the dynamic class, the object can be obtained / lost at runtime. It allows the temporary role that the class represents the object. Although multiple classifications meet common logical relationships, it has caused difficulties to programming, and common language does not support multiple classifications.
