Issue 18032: Pg. 613, Clause 17.3.4: Notation (uml25-ftf)
Source: Lockheed Martin (Mr. Lenny Delligatti, lenny_delligatti2(at)omg.org)
Nature: Uncategorized Issue
Severity: Significant
Summary: Title:  Incomplete grammar for “<lifelineident>” in Clause 17.3.4
Summary:  This clause specifies the following grammar for “<lifelineident>”:  

<lifelineident> ::= ([<connectable-element-name>[‘[‘<selector>‘]’]] [: <class_name>][decomposition]) | ‘self’

“<class_name>”, however, is too restrictive for the type.  A lifeline may represent an instance of an Actor, not just an instance of a Class.  In fact, the metamodel, as written, allows a lifeline to represent an instance of any type of BehavioredClassifier.  Here are the relationships:

A lifeline represents 0..1 instances of ConnectableElement.
ConnectableElement is a type of TypedElement.
TypedElement has an association with 0..1 instances of Type.
BehavioredClassifier is a type of Classifier, which is a type of Type.
BehavioredClassifier has 4 specializations:  Class, Actor, UseCase, and Collaboration.

Therefore, any of these 4 subtypes may serve as the type of the connectable element that a lifeline represents.  But only 2 of these make sense in the context of a lifeline:  Class and Actor.  So there are really 2 problems that need to be resolved.  Recommendations provided below

Proposed Resolution:  
1)  The grammar for “<lifelineident>” needs to be expanded to allow for an actor to serve as the type of the connectable element that a lifeline represents, and
2)  A constraint needs to be introduced to allow only two of the four subtypes of BehavioredClassifier to serve as the type of the connectable element that a lifeline represents.

Resolution: 
Revised Text: 
Actions taken:
September 27, 2012: received issue
Discussion: 

End of Annotations:=====
s is issue # 18032

Problem: 17.014

Severity:  Significant
Nature:  Grammar
Location:  Pg. 613, Clause 17.3.4: Notation
Title:  Incomplete grammar for .<lifelineident>. in Clause 17.3.4
Summary:  This clause specifies the following grammar for .<lifelineident>.:  
 
<lifelineident> ::= ([<connectable-element-name>[.[.<selector>.].]] [: <class_name>][decomposition]) | .self.
 
.<class_name>., however, is too restrictive for the type.  A lifeline may represent an instance of an Actor, not just an instance of a Class.  In fact, the metamodel, as written, allows a lifeline to represent an instance of any type of BehavioredClassifier.  Here are the relationships:
 
A lifeline represents 0..1 instances of ConnectableElement.
ConnectableElement is a type of TypedElement.
TypedElement has an association with 0..1 instances of Type.
BehavioredClassifier is a type of Classifier, which is a type of Type.
BehavioredClassifier has 4 specializations:  Class, Actor, UseCase, and Collaboration.
 
Therefore, any of these 4 subtypes may serve as the type of the connectable element that a lifeline represents.  But only 2 of these make sense in the context of a lifeline:  Class and Actor.  So there are really 2 problems that need to be resolved.  Recommendations provided below
 
Proposed Resolution:  
1)  The grammar for .<lifelineident>. needs to be expanded to allow for an actor to serve as the type of the connectable element that a lifeline represents, and
2)  A constraint needs to be introduced to allow only two of the four subtypes of BehavioredClassifier to serve as the type of the connectable element that a lifeline represents.
 
Source:         Lenny Delligatti