Home » Modeling » Papyrus » What does this mean in a sequence diagram?
What does this mean in a sequence diagram? [message #1022188] |
Thu, 21 March 2013 12:48 |
Joost Kraaijeveld Messages: 273 Registered: July 2009 |
Senior Member |
|
|
Hi,
I have a sequence diagram that, upon model validation, gives several
errors with as text:
"The required feature 'action' of '<Action Execution Specification>
ActionExecSpec' must be set".
And indeed, when I look at the properties of the lifelines, I have for
Action "undefined". Clicking on "+" gives me a choice between many
"Actions". I do not know which one to choose and why. I thought that an
Action Execution Specification basically executed the function with the
corresponding message?
TIA
--
Groeten,
Joost Kraaijeveld
Askesis B.V.
Molukkenstraat 14
6524NB Nijmegen
tel: 024-3888063 / 06-51855277
fax: 024-3608416
web: www.askesis.nl
Cheers,
Joost
|
|
| | |
Re: What does this mean in a sequence diagram? [message #1027598 is a reply to message #1022259] |
Wed, 27 March 2013 06:50 |
Marc-Florian Wendland Messages: 83 Registered: January 2013 |
Member |
|
|
Hi Joost,
>I thought that an
>Action Execution Specification basically executed the function with the
>corresponding message?
That's a common misunderstanding, because these rectangles are very often
misused to illustrate the execution duration of the called operation. The
real semantics of these ExecutionSpecifications is that after the message
has been sent and the invocation has been requested, an additional Behavior
or Action can be executed. Unfortunately, Papyrus always creates these
ExecutionSpecification rectangles always for synchronous calls.
I always work without these rectangles since they do not provide any added
value to me, because I do not want to call any additional thing. However,
as a matter of fact ActionExecutionSpecification requires an Action that is
contained in the surrounding Interaction (take a look into the metamodel). I
always link an OpaqueAction to ActionExecutionSpecification, or use
BehaviorExecutionSpecification as GoodGuy wrote. It is strange, but
BehaviorExecutionSpecification does not force you to specify a behavior, so
in order to get syntactically correct models and to misuse those rectangles
as execution duration indicator, go for BehaviorExecutionSpecification.
HTH
Marc-Florian
"Joost Kraaijeveld" wrote in message news:kif7dr$1m7$1@xxxxxxxxe.org...
Hi,
Thanks for the answer. But...
On 21-03-13 14:41, Goood Guy wrote:
> Use a CallOperationAction or alternatively use a
> BehaviorExecutionSpecification instead of an ActionExecutionSpecification.
I only have 1 class diagram, 1 application composite diagram and 1
sequence diagram. Whatever I choose as an action, Papyrus forces me
choose something, and it only allows me to choose the interaction1 thing
that contains the sequence diagram. Is that correct? And if so, what am
I specifying by doing so?
TIA
Joost
|
|
|
Re: What does this mean in a sequence diagram? [message #1028389 is a reply to message #1027598] |
Thu, 28 March 2013 07:39 |
Joost Kraaijeveld Messages: 273 Registered: July 2009 |
Senior Member |
|
|
Hi Florian,
On 27-03-13 07:50, Marc-Florian Wendland wrote:
> That's a common misunderstanding, because these rectangles are very
> often misused to illustrate the execution duration of the called
> operation. The real semantics of these ExecutionSpecifications is that
> after the message has been sent and the invocation has been requested,
> an additional Behavior or Action can be executed.
OK, thanks. I did not know that. But that explains why the message arrow
does not stick to the start of the rectangle. Mystery solved.
> Unfortunately, Papyrus
> always creates these ExecutionSpecification rectangles always for
> synchronous calls.
I even have the impression that a synchronous message *must* have its
origin in an InteractionCompartimenet? Is that correct and is that a
feature of the tool or of the specification, and if the latter, could
you give a hint of the chapter and verse?
TIA
Joost
Cheers,
Joost
|
|
|
Re: What does this mean in a sequence diagram? [message #1029254 is a reply to message #1028389] |
Fri, 29 March 2013 11:38 |
Marc-Florian Wendland Messages: 83 Registered: January 2013 |
Member |
|
|
Hi Joost,
>I even have the impression that a synchronous message *must* have its
>origin in an InteractionCompartimenet? Is that correct and is that a
>feature of the tool or of the specification, and if the latter, could
>you give a hint of the chapter and verse?
This is tool-specific.
Regards,
Marc-Florian
"Joost Kraaijeveld" wrote in message news:5153F3CE.5090905@xxxxxxxx...
Hi Florian,
On 27-03-13 07:50, Marc-Florian Wendland wrote:
> That's a common misunderstanding, because these rectangles are very
> often misused to illustrate the execution duration of the called
> operation. The real semantics of these ExecutionSpecifications is that
> after the message has been sent and the invocation has been requested,
> an additional Behavior or Action can be executed.
OK, thanks. I did not know that. But that explains why the message arrow
does not stick to the start of the rectangle. Mystery solved.
> Unfortunately, Papyrus
> always creates these ExecutionSpecification rectangles always for
> synchronous calls.
I even have the impression that a synchronous message *must* have its
origin in an InteractionCompartimenet? Is that correct and is that a
feature of the tool or of the specification, and if the latter, could
you give a hint of the chapter and verse?
TIA
Joost
|
|
|
Goto Forum:
Current Time: Fri Apr 19 19:34:16 GMT 2024
Powered by FUDForum. Page generated in 0.03841 seconds
|