Issue 2336:   extension to the notation for a transition (uml2-superstructure-ftf)
Source:  (, )
Nature:  Enhancement
Severity:  Significant
Summary: Summary: I would like to make an appeal for an extension to the notation for a transition 
 to allow its effect to be specified declaratively rather than only imperatively by 
 means of an action sequence, e.g. 
 
 e() / [p] 
 
 While I realize there are ways to work around this (e.g. by writing "e() / pTrue()" 
 where the query pTrue() has the postcondition "result = p and in targetState"), I
 think the issues are readability and ease of use. 
 
 

Resolution: 
Revised Text: This is an issues (enhancement) submitted to UML 1.4 RTF, which is addressed to a
large extent by protocol state machines syntax in UML 2.0. Therefore the
recommendation is to close the issue.
Actions taken:
January 22, 1999: received issue
March 9, 2005: closed issue
Discussion: 
This is a minor modeling enhancement to UML that would result in a significant change to the metamodel of state machine in the standard. However, it may be useful to consider it for UML 2.0.
End of Annotations:=====
Date: Fri, 22 Jan 1999 15:06:56 -0500 (EST)
From: www
To: juergen@omg.org, web-incoming@omg.org
Subject: Issue Report

Name: 
Company: 
mailFrom: 
Specification:
The syntax for a simple transition would change similarly to permit a
declared effect
following the action-expression.

Taking this one step further, a declared effect could appear anyplace
an action-expression
 could in the specification of the state machine, e.g. following a
"do:" in a state as well.

Gary Daugherty
gwdaughe@collins.rockwell.com
319-295-4065
submit: Submit Issue Report

 Gary Daugherty
Section: 9
Formal #: ad/97-08-05
Version: 1.1
Revision_Date: 9/1/97
Page: 105, 111
Nature: Enhancement
Severity: Significant
full_desc: I would like to make an appeal for an extension to the
notation for a transition 
to allow its effect to be specified declaratively rather than only
imperatively by 
means of an action sequence, e.g. 

e() / [p] 

While I realize there are ways to work around this (e.g. by writing
"e() / pTrue()" 
where the query pTrue() has the postcondition "result = p and in
targetState"), I
think the issues are readability and ease of use. 

We want the customer to be able to look at the state diagram alone in
order to 
understand the effects of the transitions and we want the notation to
be both simple 
and direct. 

This is the same notation used by Cook and Daniels in Syntropy and by
myself in my 
papers the unification of the class and state machine models.

The new syntax for an internal transition would be:


UML 2.0 Superstructure FTF
Resolution proposals for Statemachines / Ballot 10
OMG Issue No: 2336
Title: extension to the notation for a transition
Source:
Summary:
I would like to make an appeal for an extension to the notation for a transition 
 to allow its effect to be specified declaratively rather than only imperatively by 
 means of an action sequence, e.g. 
 
 e() / [p] 
 
 While I realize there are ways to work around this (e.g. by writing "e() / pTrue()" 
 where the query pTrue() has the postcondition "result = p and in targetState"), I
 think the issues are readability and ease of use. 
Discussion:
This is an issues (enhancement) submitted to UML 1.4 RTF, which is addressed to a large extent by protocol state machines syntax in UML 2.0. Therefore the recommendation is to close the issue.
Disposition: Closed, no change (resolved in UML 2.0)

event-name argument-list [