|
|
|
|
Re: Defining Component interactions in Sequence Diagram [message #1488313 is a reply to message #1486766] |
Wed, 26 November 2014 13:55 |
|
Hi,
Component0 can also inherit operations from supertypes
(generalizations), but realized interfaces are not supertypes and their
features are not inherited. A classifier realizing an interface is
responsible for providing some appropriate realizations of the
interface's features, and the form that such realizations take is a
semantic variation point (I think). FWIW, I recall some version of the
UML specification giving an example of a model in a Java-based system
in which interface properties are realized in a class by getter and
setter operations.
HTH,
Christian
On 2014-11-25 09:18:42 +0000, Grischa Liebel said:
> Hi Camille,
>
> my idea was to define a component diagram according to the attached
> image (compDiag.png). Then, I want to define two properties c0 and c1
> of type Component0 and Component1. These, I would then like to use in a
> sequence diagram and define an interaction based on the interface I
> defined in the component diagram.
> However, Papyrus knows only the operations that Component1 owns, not
> the ones that it realises through an interface. I.e. in
> sequenceDiag.png, I can't choose op() from the existing operations. For
> that, I would have to use a property of type IProvides instead.
> Is this intended?
> <image>
> <image>
|
|
|
Powered by
FUDForum. Page generated in 0.03623 seconds