[news.eclipse.tools.uml2] Re: Action Semantics support

["Antonio Carrasco Valero" <acv@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>]

Brilliant !

  Yep! you are right. No human textual language for it! nobody would care, 
we already have those!
  It seems to be a Platform Independent metamodel for imperative object 
languages ...




 - Normally, I use it when I don´t want to go into full algorithmic detaul 
at all,
   and only use a few of the Actions,
   and not with the goal of producing a precise specification (a-la- pUML 
way)
   so I can not emit an opinion about ambiguity.

   Actions work for me, because the Activity-Action composition
   and its architectural isomorphism with EncapsulatedClassifiers 
LogicalComponents with ports.
   So that we can explain as Activities and Actions,
   the choreographed behavior of structural assemblies of LogicalComponents.

    (A usage inspired in the Component Collaboration Architecture and 
Business Process Profile)

http://www.omg.org/technology/documents/profile_catalog.htm#UML_for_EDOC
specs in
http://www.omg.org/cgi-bin/doc?formal/2004-02-01
http://www.omg.org/cgi-bin/doc?formal/2004-02-05


Thanks,
Antonio


"James Bruck" <jbruck@xxxxxxxxxx> wrote in message 
news:eh5n38$gm6$1@xxxxxxxxxxxxxxxxxxxx
> Hi Antonio,
>
> To explain a bit better ...
>
> I can see the ambiguity to the question about "Action Semantics".    I 
> guess
> the answer really depends on what was originally intended by the original
> question.
> The UML metamodel does support actions and can be said to define "Action
> Semantics" in that it is a way to specify actions.   In that case,
> opensource UML does support "Action Semantics".
>
> However, from what I understand the term "Action Semantics" usually refers
> to an OCL type language to bridge the gap between ambiguity in the UML
> metamodel and fully specified actions.
>
> The UML specification of Actions has come a long way in the latest version
> so I am not certain how much of the original ambiguity as existing in 
> UML1.x
> still exists currently.
>
> I hope that clarifies.
>
> - James.
>
>
>
> "James Bruck" <jbruck@xxxxxxxxxx> wrote in message
> news:eh0ua3$fsp$1@xxxxxxxxxxxxxxxxxxxx
>> Hi Antonio.
>>
>> Opensource UML implements action metaclasses as specified in the spec but
>> there is no facility to specify Action Semantics language ( kinda like
>> OCL ).
>> What you are referring to is not Action Semantics but the existance of
>> metaclasses.
>>
>> Cheers,
>>
>> - James.
>>
>>
>>
>> "Antonio Carrasco Valero" <acv@xxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote in
>> message news:eh0n9m$ofd$1@xxxxxxxxxxxxxxxxxxxx
>> > James,
>> >
>> > I believe that the answer to Andrey is Yes, we do have Action 
>> > Semantics,
>> > because:
>> >
>> > I observe that Eclipse UML2 implementation includes the classes
> originally
>> > found in the
>> > Actions
>> > ..BasicActions
>> > ..IntermediateActions
>> > ..CompleteActions
>> > ..StructuredActions
>> >
>> > which certainly are "...to some language neutral facility for 
>> > specifying
>> > detail-level behavior in the context of a UML model..."
>> >
>> > I know Action Language is not a favourite around some in the company
>> > (...C++ with boxes and lines ...), but I can see it perfectly in the
> UML2
>> > sources.
>> >
>> > Thanks,
>> > Antonio Carrasco Valero
>> > Model Driven Development sl
>> >
>> >
>> >
>> > "James Bruck" <jbruck@xxxxxxxxxx> wrote in message
>> > news:egitpu$60q$1@xxxxxxxxxxxxxxxxxxxx
>> > > Hi Andrey,
>> > >
>> > > If you are referring to some language neutral facility for specifying
>> > > detail-level behavior in the context of a UML model the answer is 
>> > > that
>> it
>> > > is
>> > > not directly supported by opensource UML currently.
>> > >
>> > > - James.
>> > >
>> > >
>> > > "Andrey Sadovykh (SOFTEAM)" <andrey.sadovykh@xxxxxxxxxx> wrote in
>> message
>> > > news:8c0c47a264da13098152b958336a1c18$1@xxxxxxxxxxxxxxxxxx
>> > >> Hello,
>> > >>
>> > >> I would like to know whether Action Semantics is implementation in
> EMF
>> > >> UML2 metamodel. If yes, on what level?
>> > >>
>> > >> regads,
>> > >> andrey
>> > >>
>> > >
>> > >
>> >
>> >
>>
>>
>
>