Datasets:
bongsoo
/
moco_eval

Dataset card Data Studio Files
xet
Community
text
stringlengths
16
313
A complete description ofthe OCL syntax and semantics can be found in [Warmer-99].
The following se-lected summary contains the most useful OCL syntax for creating navigation ex-pressions and Boolean conditions.
Syntax for some common navigation expressions is shown below.
The left-most element must be an expression for an object ora set of objects.
The expressions are meant to work on sets of values when applica-ble.
For more details and syntax, see the OCL description.
The result is the value of the attribute or the related object(s).
The result is the related object selected by the qualifier.
Note that this syntax is applicable to array indexing as a form of qualification.
The result is the property of the set.
set -> select ( boolean-expression )
The result is the subset of objects in the set for which the Boolean expression is true.
The set of pilots who have enough training hourscompany.
An operation is a specification of a transformation or query that an object may becalled to execute.
It has a name and a list of parameters.
A method is a procedurethat implements an operation.
It has an algorithm or procedure description.
See also call, call event, method.
An operation specifies a transformation on the state of the target object (and pos-sibly the state of the rest of the system reachable from the target object) or a querythat returns a value to the caller of the operation.
An operation may be imple-mented as a a method or as a call event that causes a transition in the state ma-chine of an active object.
An operation is invoked by a call, which suspends thecaller until the execution of the operation is complete, after which the caller re-sumes control beyond the point of the call, receiving a return value if one is sup-plied by the operation.
An operation is declared in a class.
The declaration is inherited by the descen-dants of the class.
If another declaration has the same “matching signature,” it isthe same operation.
An implementation may specify a rule for matching signa-tures to test for conflict, but by default, it includes the name of the operation andthe classes (but not the names or directions) of the parameters, not including re-turn parameters.
The same operation can appear in a descendant class.
In thatcase, it is treated as a repetition of the inherited declaration and ignored.
The pur-pose is to permit an operation to be declared multiple times in classes that are de-veloped in different packages, using name matching.
It represents the governing declaration of the operation that isinherited by the others.
A method is the implementation of an operation (it may also be implementedby a call event).
If an operation is declared in a class without the abstract property,then it has a method definition in the class.
Otherwise, the operation may be ab-stract (and there is no method), or it may be concrete with an inherited method.
An operation has the following main constituents.
Possible values aresequential Callers must coordinate so that only one call to anobject (on any sequential operation) may execute atonce.
If concurrent calls occur, then the semanticsand integrity of the system cannot be guaranteed.
The others are blocked until the execution ofthe first operation is complete.
It is the responsibil-ity of the modeler to ensure that deadlocks do notoccur because of simultaneous blocks.
Guarded op-erations must perform correctly (or block them-selves) in the case of a simultaneous sequentialoperation, or guarded semantics cannot beclaimed.
All of them may proceed concurrently withcorrect semantics.
Concurrent operations must bedesigned so that they perform correctly in the caseof a concurrent, sequential, or guarded operationon the same object.
Otherwise, concurrent seman-tics cannot be claimed.
If false, the current implementation isinherited unchanged by all descendants.
If false, it may alter the state of the system, but achange is not guaranteed.
with the list of parameter types (not including parameternames or return types), is called the matching signatureof the operation.
The matching signature must be uniquewithin the class and its ancestors.
If there is a duplication,it is taken as a redeclaration of the operation, which mustmatch completely.
If they match, all but the operationdeclaration in the highest ancestor are ignored.
If they donot match, the model is ill formed.
return type A list of the types of the values returned by a call of theoperation, if any.
then this property has the value null.
Note that many lan-guages do not support multiple return values, but itremains a valid modeling concept that can be imple-mented in various ways, such as by treating one or moreof the parameters as output values.
Possible values areinstance The operation may be applied to individual objects.
The format of the specification is not prescribedby UML and can take various forms.
A method has the same constituents as an operation.
In addition, it may have oneor more ofbehavior An optional state machine describing the implementationof the method.
Usually, this would be expressed in a program-ming language, although a natural language expression ispossible for informal specifications.
Generally, this valuewould not be supplied if the state machine is supplied.
A call event has the same constituents as an operation.
The implementation of theoperation must be specified by one or more transitions that have the call event as atrigger.
An operation is shown as a text string that can be parsed into properties of the op-eration.
The stereotype, visibility, return-type-expression, and property string are op-tional (together with their delimiters).
The parameter list may be empty.
Figure 13-137 shows some typical operations.
A string that is the name of the operation (not including parameters).
A string containing a comma-separated list of names of classifiers(classes, data types, or interfaces).
The type string follows a colon (:) that followsthe parameter list of the operation.
Some, but not all, pro-gramming languages support multiple return values.
The visibility is shown as one of the punctuation marks ‘+’, ‘#’, or ‘-’,representing public, protected, or private.
Alternately, visibility can be shown as akeyword within the property string (for example, {visibility=private}).
This formmust be used for user-defined or language-dependent choices.
An operation and a method are declared using the same syntax.
The top-most appearance of an operation signature within a generalization hierarchy is thedeclaration of an operation.
Identical signatures in descendant classes are redun-dant declarations of the operation, but these may be useful for declaring methodsor for declaring operations when the classes are developed separately.
If an opera-tion declaration has the abstract property (notated by operation name in italics orthe keyword abstract), then there is no method corresponding to the declaration.
Otherwise, the declaration represents both an operation declaration and a methodimplementing it.
In matching operations and methods, the name of the operation and the or-dered list of parameter types are used, not including return parameters.
If the re-maining properties are inconsistent (for example, an in-parameter is matched toan out-parameter), then there is a conflict and the model is ill formed.
If two identical operation declarations have no common ancestor operationdeclaration, yet are inherited by a common class, then the model is ill formed.
The body of a method may be shown as a string within a note at-tached to an operation declaration.
Otherwise, it should be normal text if it is just a natural-language de-scription of the behavior (a comment).
The connection of a method declarationand its state machine or collaboration has no visual representation, but it wouldgenerally be represented by a hyperlink within an editing tool.
An expression describing the effects of performing the operation.
This may be stated in various ways, including text, before-after conditions, and in-variants.
In any case, the specification should be expressed in terms of the observ-able effects of the operation on the state of the system, not in terms of theexecution algorithm.
The algorithm is the business of the method.
The choice true may also be shown by the keyword query.
Theabsence of an explicit choice indicates the choice false—that is, the operation mayalter the system state (but it does not guarantee to alter it).
Theabsence of an explicit choice indicates the choice true—that is, overridable.
An instance-scope operation is indicated by not underlining the operationstring.
A class-scope operation is indicated by underlining the name string.
The choice is shown by a property string of the form concur-rency=value, where the value is one of sequential, guarded, or concurrent.
To indicate that a class accepts a signal, the keyword «signal» is placed infront of an operation declaration within the list of operations.
The declaration may not have a return type.
The re-sponse of the object to the reception of the signal is shown with a state machine.
Among other uses, this notation can show the response of objects of a class to er-ror conditions and exceptions, which should be modeled as signals.
The argument list and return type may be suppressed (together, not separately).
A tool may show the visibility indication in a different way, such as by using a spe-cial icon or by sorting the elements by group.
The syntax of the operation signature string can be that of a particular program-ming language, such as C++ or Smalltalk.
Operation names typically begin with a lowercase letter.
Operation names are shown in plain face.