Sparx Systems Forum
Enterprise Architect => Suggestions and Requests => Topic started by: UlrichF on June 16, 2015, 08:01:45 pm
-
If stereotypes are being applied to ArchiMate elements, the stereotypes do not show up on diagrams (independent whether you choose rectangle or non-rectangle notation).
Stereotypes (as well as tags) should be shown for ArchiMate elements.
-
I believe the reason they aren't is because ArchiMate doesn't have a concept of stereotypes.
-
I believe the reason they aren't is because ArchiMate doesn't have a concept of stereotypes.
Fair comment.
However, ArchiMate itself intends "to provide a means to allow extensions of the core language that are tailored towards such specific domains or applications, without burdening the core with a lot of additional concepts and notations ... The remainder of this section is devoted to a number of possible extension mechanisms that, in addition to the core, are or can become part of the ArchiMate language." It mentions profiles and UML stereotypes as means of achieving language extensibility. (http://pubs.opengroup.org/architecture/archimate2-doc/chap09.html).
-
Create a profile or MDG that extends the Archimate like I have here
https://onedrive.live.com/redir?resid=D323B9F737BF685E!18705&authkey=!AE6b76Bo28Az4zM&ithint=file%2cxml
-
I believe the reason they aren't is because ArchiMate doesn't have a concept of stereotypes.
Simon is quite right, and I think it is instructive to consider why...
In my experience both the standards organisations and the tool makers have (over the decades - and with some sympathy [the frog in the heating water problem]) conflated the properties of type, metatype and stereotype.
EA's use of multiple stereotypes - particularly when allied to shape scripts is less than optimal. However, the definition of a stereotype and how we apply them is a bit confusing.
The concept of using type, stereotype combinations to define new metatypes (as implemented in EA) goes part of the way towards solving the problem.
What, I believe, is required is the ability to define (real) new types (the kinds of things that go into the type column of t_object - as metatypes.
Anyway, in our extensions to ArchiMate, we have come to the view that we will (hopefully) ONLY use stereotypes to define metatypes - as Sparx have done. We'll use other mechanisms (such as Tagged Values) to distinguish between variants of a metatype.
HTH,
Paolo
-
And if you look across the Archimate documentation and the Open Group white papers you will see inconsistency in how they handle language extension.