Home » Modeling » UML2 » Combined Fragments
| |
| Re: Combined Fragments [message #474267 is a reply to message #474264] |
Tue, 17 July 2007 08:39   |
 Andrew Carton
Messages: 104
Registered: July 2009 |
Senior Member |
|
|
Thanks James, that made it a bit clearer for me. I must be doing
something incorrectly with Magicdraw then, I will contact their support.
Cheers,
Andrew.
Ar 16/07/2007 20:35, Scríobh James Bruck:
> Hi Andrew,
>
> The containing interaction would have a list of fragments that should
> include the alt combined fragment ( looking at your picture ).
> In your case you should be seeing a InteractionOperand owned by your alt
> combined fragment.
> That interaction operand would own behavior execution specification and
> message occurrence specifications ( corresponding to the end points and
> execution of your message blah() )
> Your message blah() would be owned by the interaction that also owns the alt
> combined fragment.
>
> I don't think gates are necessary for what you are looking for.
>
> - James.
>
>
> "Andrew Carton" <cartona@cs.tcd.ie> wrote in message
> news:f7fl0d$8ti$1@build.eclipse.org...
>> Hi,
>>
>> I am having a bit of a problem understanding Combined Fragments in UML
>> 2.1. I am using Magicdraw 12.5 to draw my sequence diagram and an
>> alternate (see attached pic) and then export to EMF (attached too).
>>
>>
>> I don't see how the different alternate fragments are linked in with the
>> container seq. diagram. Perhaps I am missing something, but there is
>> no reference in the combined fragment to which Messages belong to it. In
>> the uml 2.1 superspec, it describes gates being used as interfaces to
>> and from the combined fragments. Magicdraw, as I am aware doesn't
>> support gates. So my question is, are gates necessary to link in
>> combinedfragments to the container interaction?
>>
>> I'd appreciate it if anyone could help me shed some light on this!
>>
>> Thanks,
>> Andrew Carton.
>>
>>
>>
>>
>
>
> ------------------------------------------------------------ ----------------
> ----
>
>
>
>
>
|
|
| |
| Re: Combined Fragments [message #474269 is a reply to message #474268] |
Tue, 17 July 2007 14:17 |
Andrew Carton
Messages: 104
Registered: July 2009 |
Senior Member |
|
|
Hi James,
I think I understand now the combined fragments and the migration guide
does give some details on the modifications to the spec, as I heard
there was a bit of a problem with them in UML 2.0.
I understand that there was a mistake or two in the diagram but I
fiddled about with magicdraw and corrected it from the last email you
sent and contacted their support. I also tried out the EMF 2.x export
instead of the EMF 1.x, like you advised. Unfortunately, they informed
me that the interaction operands and the corresponding interaction
fragments aren't linked in the EMF UML2 export.
Thanks for your help though, its much appreciated!
Andrew Carton.
Ar 17/07/2007 13:46, Scríobh James Bruck:
> Hi Andrew,
>
> I think the key problem is that the operand is not owning the end points of
> the messages ( MessageOccurrenceSpecifications ). That's the main way to
> tell that the message is owned by the alt fragment.
> Also, I noticed that this is a .uml2 file indicating this is an older
> version. Have a look at the migration document for an explanation of how
> Sequence diagrams have changed since the older revision of the spec.
> From looking at the diagram, there also seems to be a few issues.
>
> 1. A circle in the upper corner of the combined fragment - not sure what
> that is.
> 2. Execution specification weaving in and out of the combined frag.
> 3. Missing execution specification for blah().
> 4. Incorrectly placed message end points.
>
> Cheers.
>
> - James.
>
> "Andrew Carton" <cartona@cs.tcd.ie> wrote in message
> news:f7hv8d$ju5$1@build.eclipse.org...
>> Thanks James, that made it a bit clearer for me. I must be doing something
>> incorrectly with Magicdraw then, I will contact their support.
>>
>> Cheers,
>> Andrew.
>>
>> Ar 16/07/2007 20:35, Scríobh James Bruck:
>>> Hi Andrew,
>>>
>>> The containing interaction would have a list of fragments that should
>>> include the alt combined fragment ( looking at your picture ).
>>> In your case you should be seeing a InteractionOperand owned by your alt
>>> combined fragment.
>>> That interaction operand would own behavior execution specification and
>>> message occurrence specifications ( corresponding to the end points and
>>> execution of your message blah() )
>>> Your message blah() would be owned by the interaction that also owns the
>>> alt
>>> combined fragment.
>>>
>>> I don't think gates are necessary for what you are looking for.
>>>
>>> - James.
>>>
>>>
>>> "Andrew Carton" <cartona@cs.tcd.ie> wrote in message
>>> news:f7fl0d$8ti$1@build.eclipse.org...
>>>> Hi,
>>>>
>>>> I am having a bit of a problem understanding Combined Fragments in UML
>>>> 2.1. I am using Magicdraw 12.5 to draw my sequence diagram and an
>>>> alternate (see attached pic) and then export to EMF (attached too).
>>>>
>>>>
>>>> I don't see how the different alternate fragments are linked in with the
>>>> container seq. diagram. Perhaps I am missing something, but there is
>>>> no reference in the combined fragment to which Messages belong to it. In
>>>> the uml 2.1 superspec, it describes gates being used as interfaces to
>>>> and from the combined fragments. Magicdraw, as I am aware doesn't
>>>> support gates. So my question is, are gates necessary to link in
>>>> combinedfragments to the container interaction?
>>>>
>>>> I'd appreciate it if anyone could help me shed some light on this!
>>>>
>>>> Thanks,
>>>> Andrew Carton.
>>>>
>>>>
>>>>
>>>>
>>>
>>> ------------------------------------------------------------ ----------------
>>> ----
>>>
>>>
>>>
>>>
>
|
|
| |
| Re: Combined Fragments [message #623744 is a reply to message #474264] |
Tue, 17 July 2007 08:39 |
Andrew Carton
Messages: 104
Registered: July 2009 |
Senior Member |
|
|
Thanks James, that made it a bit clearer for me. I must be doing
something incorrectly with Magicdraw then, I will contact their support.
Cheers,
Andrew.
Ar 16/07/2007 20:35, Scríobh James Bruck:
> Hi Andrew,
>
> The containing interaction would have a list of fragments that should
> include the alt combined fragment ( looking at your picture ).
> In your case you should be seeing a InteractionOperand owned by your alt
> combined fragment.
> That interaction operand would own behavior execution specification and
> message occurrence specifications ( corresponding to the end points and
> execution of your message blah() )
> Your message blah() would be owned by the interaction that also owns the alt
> combined fragment.
>
> I don't think gates are necessary for what you are looking for.
>
> - James.
>
>
> "Andrew Carton" <cartona@cs.tcd.ie> wrote in message
> news:f7fl0d$8ti$1@build.eclipse.org...
>> Hi,
>>
>> I am having a bit of a problem understanding Combined Fragments in UML
>> 2.1. I am using Magicdraw 12.5 to draw my sequence diagram and an
>> alternate (see attached pic) and then export to EMF (attached too).
>>
>>
>> I don't see how the different alternate fragments are linked in with the
>> container seq. diagram. Perhaps I am missing something, but there is
>> no reference in the combined fragment to which Messages belong to it. In
>> the uml 2.1 superspec, it describes gates being used as interfaces to
>> and from the combined fragments. Magicdraw, as I am aware doesn't
>> support gates. So my question is, are gates necessary to link in
>> combinedfragments to the container interaction?
>>
>> I'd appreciate it if anyone could help me shed some light on this!
>>
>> Thanks,
>> Andrew Carton.
>>
>>
>>
>>
>
>
> ------------------------------------------------------------ ----------------
> ----
>
>
>
>
>
|
|
| |
| Re: Combined Fragments [message #623746 is a reply to message #474268] |
Tue, 17 July 2007 14:17 |
Andrew Carton
Messages: 104
Registered: July 2009 |
Senior Member |
|
|
Hi James,
I think I understand now the combined fragments and the migration guide
does give some details on the modifications to the spec, as I heard
there was a bit of a problem with them in UML 2.0.
I understand that there was a mistake or two in the diagram but I
fiddled about with magicdraw and corrected it from the last email you
sent and contacted their support. I also tried out the EMF 2.x export
instead of the EMF 1.x, like you advised. Unfortunately, they informed
me that the interaction operands and the corresponding interaction
fragments aren't linked in the EMF UML2 export.
Thanks for your help though, its much appreciated!
Andrew Carton.
Ar 17/07/2007 13:46, Scríobh James Bruck:
> Hi Andrew,
>
> I think the key problem is that the operand is not owning the end points of
> the messages ( MessageOccurrenceSpecifications ). That's the main way to
> tell that the message is owned by the alt fragment.
> Also, I noticed that this is a .uml2 file indicating this is an older
> version. Have a look at the migration document for an explanation of how
> Sequence diagrams have changed since the older revision of the spec.
> From looking at the diagram, there also seems to be a few issues.
>
> 1. A circle in the upper corner of the combined fragment - not sure what
> that is.
> 2. Execution specification weaving in and out of the combined frag.
> 3. Missing execution specification for blah().
> 4. Incorrectly placed message end points.
>
> Cheers.
>
> - James.
>
> "Andrew Carton" <cartona@cs.tcd.ie> wrote in message
> news:f7hv8d$ju5$1@build.eclipse.org...
>> Thanks James, that made it a bit clearer for me. I must be doing something
>> incorrectly with Magicdraw then, I will contact their support.
>>
>> Cheers,
>> Andrew.
>>
>> Ar 16/07/2007 20:35, Scríobh James Bruck:
>>> Hi Andrew,
>>>
>>> The containing interaction would have a list of fragments that should
>>> include the alt combined fragment ( looking at your picture ).
>>> In your case you should be seeing a InteractionOperand owned by your alt
>>> combined fragment.
>>> That interaction operand would own behavior execution specification and
>>> message occurrence specifications ( corresponding to the end points and
>>> execution of your message blah() )
>>> Your message blah() would be owned by the interaction that also owns the
>>> alt
>>> combined fragment.
>>>
>>> I don't think gates are necessary for what you are looking for.
>>>
>>> - James.
>>>
>>>
>>> "Andrew Carton" <cartona@cs.tcd.ie> wrote in message
>>> news:f7fl0d$8ti$1@build.eclipse.org...
>>>> Hi,
>>>>
>>>> I am having a bit of a problem understanding Combined Fragments in UML
>>>> 2.1. I am using Magicdraw 12.5 to draw my sequence diagram and an
>>>> alternate (see attached pic) and then export to EMF (attached too).
>>>>
>>>>
>>>> I don't see how the different alternate fragments are linked in with the
>>>> container seq. diagram. Perhaps I am missing something, but there is
>>>> no reference in the combined fragment to which Messages belong to it. In
>>>> the uml 2.1 superspec, it describes gates being used as interfaces to
>>>> and from the combined fragments. Magicdraw, as I am aware doesn't
>>>> support gates. So my question is, are gates necessary to link in
>>>> combinedfragments to the container interaction?
>>>>
>>>> I'd appreciate it if anyone could help me shed some light on this!
>>>>
>>>> Thanks,
>>>> Andrew Carton.
>>>>
>>>>
>>>>
>>>>
>>>
>>> ------------------------------------------------------------ ----------------
>>> ----
>>>
>>>
>>>
>>>
>
|
|
|
Goto Forum:
Current Time: Tue Apr 23 17:19:32 GMT 2024
Powered by FUDForum. Page generated in 0.04046 seconds |
] 
Search
Help
Register
Login
Home
