|
|
|
| An attribute is a string item listed in the attribute compartment of a classifier. An attribute defines an aspect or dimension of a classifier. An attribute may be assigned a value when an instance of a classifier is created. It is semantically equivalent to a classifier having a composite association with another classifier of the type specified in the attribute expression. |
|
|
|
|
| A value of bag specified for the property string of an association end indicates that the same instance can appear more than once in the collection. By default an association end is a set of instances, meaning that there cannot be duplicate elements. By setting the value to bag this default behavior is overridden and duplicates can occur. |
|
|
|
|
| A call dependency connects an operation in one class, the source to an operation in another class, the target. The relationship may be made less precise by attaching either the source or target or both ends to the classes that contain the operations. |
|
|
|
|
| A comment can contain textual information or graphical symbols that add additional information to a diagram or model. The information can take many forms from very formal, in the case of constraints, to conversational, in the case of working notes. Comments can be placed anywhere in a model or diagram, and simply by their proximity and alignment; they can be applicable to one or more elements. They may also be attached either to one or more elements by a comment attaching line. |
|
|
|
|
| Complete is a constraint added to a set of generalization relationships to indicate that all children have been specified. This means that any instance of the parent must be an instance of at least one of the children within the set. All children that are governed by the constraint must have the same discriminator. |
|
|
|
|
| A create dependency is a usage dependency that signifies that one or more of the operations on the client (tail end) create instances of the supplier classifier (arrowhead end). |
|
|
|
|
| A deep history is a kind of pseudostate that acts as a marker or placeholder. It represents the most recent active substate of the containing composite state, and the substates of that substate. It is not a state itself but a diagrammatic representation or marker for the "condition" of the composite state down to any level, at the time it was last exited. |
|
|
|
|
| A derive dependency is an abstraction dependency that indicates that the client (tail end) can always be calculated from the supplier (arrowhead). This implies that the client is strictly redundant, and may be modeled for performance reasons or ease of use. |
|
|
|
|
| Disjoint is a constraint added to a set of generalization relationships to indicate that an instance of the parent may be an instance of one and only one of the children within the set. |
|
|
|
|
| Extend is a relationship that may exist between two use cases, the base use case and the extension use case. It is said to be conditional in the sense that a set of extension points defines conditions under which the insertion will occur. The extension points are defined in a separate section of the use case. |
|
|
|
|
| An extension point is a feature of a use case and acts as a reference to a location in the (extended) use case at which the behavior of the use case may be extended by the insertion of additional behavior from another (extending) use case. A use case may have any number of extension points and these are used in conjunction with the extend relationship which contains reference to the extension points and conditions which govern whether the extending behavior will be inserted. |
|
|
|
|
| A framework is a stereotyped package that contains elements that specify an architecture or part of an architecture that can be reused. The extent of the framework may be for an entire system or may be restricted to a part of a system. |
|
|
|
|
| Include is a relationship that may exist between two use cases, a base use case and an addition use case. It means a base use case may insert behavior from an addition use case. The relationship is said to be unconditional in the sense that if the insertion occurs; all the behavior of the addition use case is unconditionally inserted. |
|
|
|
|
| Incomplete is a constraint added to a set of generalization relationships to indicate that all children have not been specified. This means that an instance of the parent is not necessarily an instance of any of the children. All children that are governed by the constraint must have the same discriminator. |
|
|
|
|
| An instantiate dependency is a usage dependency that exists between classifiers and indicates that the client classifier (tail end) creates instances of the supplier (arrowhead end). |
|
|
|
|
| An interface can be represented by a classifier (rectangular) symbol. Both this form, and the more commonly seen "lollipop" (small circle and stem) form, have the same semantics. The circular form is often used when a structural view of a system is required. When an interface has operations and these need to be displayed as part of a diagram the rectangular form can be used, the operations appear in a separate compartment. |
|
|
|
|
| multiplicity is a property of an association end that specifies the number of instances of the classifier connected at the target end (near end) that can be associated with a single instance of the classifier at the source end (far end). It is specified in the form of a string stipulating a possible infinite number governed by a lower and upper bound. |
|
|
|
|
| An operation is a string item listed in the operation compartment of a classifier. The operations define the behavior of the classifier and the work that the classifier can carry out. An operation can take a list of parameters and return a type. Its visibility can be restricted and other factors, such as concurrency, can be specified using the property-string. |
|
|
|
|
| The value ordered can be set for the ordering property of an association end. This signifies that instances of the classifiers at the target end of the association are ordered. Any operation that adds or changes the instances is responsible for maintaining the order. The multiplicity at the target end must be greater than one since a value of one or less would specify a set with one or less elements, for which ordering would be meaningless. |
|
|
|
|
| Overlapping is a constraint added to a set of generalization relationships to indicate that an instance of one child may simultaneously be an instance of another. |
|
|
|
|
| When the value of package is specified for the visibility property of an association end it signifies that any classifier that is contained in the same package as the association (or any nested sub package) may traverse the relationship and use the role name in expressions. |
|
|
|
|
| When the value of private is specified for the visibility property of an association end it signifies only the classifier connected at the source (far) end may traverse the relationship and use the role name in expressions. |
|
|
|
|
| When the value of protected is specified for the visibility property of an association end it signifies that the classifier at the source (far) end of the association and its descendants may traverse the relationship and use the role name in expressions. |
|
|
|
|
| When the value of public is specified for the visibility property of an association end it signifies that other classifiers in addition to the classifier connected at the source end of the association can traverse the relationship and use the role name in expressions. |
|
|
|
|
| A value of redefines specified for the property string indicates that the association end is redefined with the name specified. This mechanism allows an association end to be recast using a new name and potentially new properties that are more appropriate at any given level of specialization (redefinition).
|
|
|
|
|
| A refine dependency is an abstraction dependency that implies that the client (tail end) is a refinement or more evolved element. This refinement may signify different version of the same element or extensions. |
|
|
|
|
| A send dependency is a type of usage dependency that exists between an operation in a classifier and a signal. It signifies that the operation (tail end) can send the signal. (arrowhead end). |
|
|
|
|
| A value of sequence specified for the property string of an association end indicates that the collection of instances are an ordered bag. This means that the collection is ordered but can contain duplicated instances. It does not specify the basis for the order. |
|
|
|
|
| A shallow history is a kind of pseudostate that acts as marker or placeholder in a composite state. It represents the state or condition of the modeled element at the time it was last exited. It is not a state itself but a diagrammatic representation or marker for the "condition" of the composite state down to a single level, at the time it was last exited. |
|
|
|
|
| A value of subsets specified for the property string of an association end indicates that the collection does not contain all the available instances defined by the named property. This is an application of the mathematical concept of subset, which means that every instance at the association end marked as subsets, must exist in the parent property. |
|
|
|
|
| A subsystem is a grouping or package of model elements that has operations and interfaces. It has the characteristics of a package and a classifier. While a subsystem is a behavior element, it has no intrinsic behavior of it s own, the behavior is provided by the elements in the realization compartment of the subsystem. |
|
|
|
|
| A trace dependency is an abstraction dependency between a supplier and a client that indicates that one or more elements in one model represent an analogous concept in another model. Its meaning depends on the relationship between the models. |
|
|
|
|
| A value of union specified for the property string indicates that the association end is derived by the union of the subsets with the name specified. This mechanism allows the classifier at the association end marked with union to specify its instances based on the union of its subsets. This is an application of the mathematical concept of union of sets and subsets. This implies that the collection is a set and cannot be a bag since a union results in a set, which means it cannot have duplicate elements. |
|
|
|
|
| The value unordered for the ordering property signifies that instances of the classifiers at the target end of the association are not ordered. This is the default value for this property and if a value for the property is not specified then the value by default is unordered. The multiplicity at the target end must be greater than one since a value of one or less would specify a set with one or less elements, for which ordering would be meaningless. |
|
