Version: 1.2 (using KDE KDE 3.1.3KDE 1.2) Installed from: Compiled From SourcesCompiled From Sources Compiler: g++ 3.2 OS: Linux uml supports stereotypes at the ModelElement class (as per the 1.5 standard). this means that almost all uml objects (anything derived from UMLObject) should support stereotyes. this includes attributes, methods, parameters (i think), and various other objects (like relationships). also, the standard defines a ModelElement to contain a set of stereotypes. umbrello should allow the modeler to create several stereotypes for any single object. for example, consider the following c++ method defined for some class: virtual void foo() const; the modeler could define the following stereotypes for the operation object: <<c++_operation>> <<virtual>> <<const>>
This is partly a duplicate of the resolved 53381 but contains some new issues.
I spent a bit of time looking howto use in a use case diagram the <<extend>> and <<include>> dependencies between use cases. What I found (maybe wrong) is that one need to write <<include>> and <<extend>> as labels on dependencies. To me this is not the best way it could be done for same reasons: - one cannot use the dropdown selection list for stereotypes and has to manually write them every times (or copy&paste) - one has to manually write the guillemets using for example KCharSelect or using double < and > - (Most important reason) extend and include are stereotypes so I think capture it in the model as stereotypes is the right way to do it. So, concluding, having umbrello supporting stereotypes for all uml objects is also useful for use cases. Should I submit a new bug [whish] for the dropdown stereotypes selection list on use case dependencies or is this comment sufficient?
Given that the core development group does not seem to be finding time for this, IMHO it would be an ideal project for a uni (ISI?), SOC (Google Summer of Code), or other external contributors. > Should I submit a new bug [...] IMHO this is sufficient as full support of stereotypes would require something in the way of such dropdown lists for all object types.
(In reply to Antonio from comment #2) > dropdown stereotypes selection list on use case dependencies This has been implemented with version v2.13.1 (KDE Version 4.13.1) commit http://commits.kde.org/umbrello/b10cdc257063aa446098b53b4dfdbd1ca1fcb453
> to contain a set of stereotypes. With version 2.15.80 (KDE version 15.03.80) commit http://commits.kde.org/umbrello/aede266afbedf7c0983776e44a4e20108be36b63 umbrello got a refactoring of the widget displaying stereotypes in the property dialogs. With this it is much easier to extend the widget to display a set of stererotypes. What is left is: 1. Display a set of stereotypes in diagram widgets 2. save to and load set of stereotypes from an xmi file 3. undo/redo support (may be implemented with topic 2.) 4. optional copy and paste support (may be implemented with topic 2.)