MDG for connectors

[vrieg]

Hello all,

we have build an MDG based on SysML for structural elements as blocks, ports, activities,...
Now we would like to add the stereotypes for the connectors to our MDG.

Is there any proposal or guide on how the different connectors available for SysML can be extended or included to a custom MDG, such that their properties i.e. Target Aggregation "composite" for "composed of" connectors can be set automatically when dragging the connector from the toolbox?

Is it recommended to base the connectors for "Aggregation", "Composite" from their respecitve Metaclasses or should we take "Association" Metaclass instead?

Or is it even recommended to extend a SysML profile?

[qwerty]

I would say it depends on your needs/goals. Whatever you think that fits best. Any of the above are possible. Going back to basic definitions leaves space for own semantics. Taking stereotypes as base will inherit theirs.

q.

[vrieg]

for the connectors our main aim is to ensure that model transformations and a potential migration to SysML V2 is not endangeoured.

[vrieg]

I added the Metaclass UCExtend to my MDG profile and used an extension to a stereotype mystereotype. That stereotype is used within a toolbox. MyMDG::Extend(UML::UCExtend)=Extend
When I use the stereotype within a toolbox I get two stereotypes <<extend>>,MyMDG::<mystereotype>.
I think that lower case <<extend>> is already part of the metaclass. Is this a bug or is there a way how the stereotype <<extend>> can be surpressed?

[qwerty]

The case sensitivity of stereotypes is a Sparx "feature"

Pls. read https://stackoverflow.com/questions/70233399/does-the-case-of-a-stereotype-matter

q.

[vrieg]

  well this is something one definitely needs to know.

while reading through the MDG documentation I  found some more issues I dont understand. As far as I understand Aggregation is outdated and for compositon stereotype within the Toolbox, rather Association should be used to together with compositionKind:Composition at Target.
Well however i dont get an Aggregation this way, but an association with composition at target.

Composition Metaclass does not work either. Any ideas. I already scrolled the forum but I rather encoundered more questions

[qwerty]

You will find lots of entries about aggregation/composition. The UML spec took care since UML 1.5. So shared/composite aggregation is now an attribute of an association (IIRC they were formerly some sort of own connectors, but not sure). Anyhow, EA made them separate and later added the association attributes. They did not deprecate the old ones and now the users have to continuously struggle with the heritage. (I guess that old Generaliz/sation jerk still hops around and fools people; just to be backward compatible.)

q.

[vrieg]

I meanwhile fixed most of the stereotypes.

However 2 errands are appearing in the MDG toolbox.

One is for
Metaclass: UCExtend
_MeaningForwards = UseCase from
_MeaningBackwards = UseCase to
Extension to
Stereotype: extend
Referenced in Toolbox as MyMDG::extend(UML::UCExtend)=Extend

--> For UCinclude doing the same way is working

Second one is
Metaclass: Nesting
_MeaningForwards = owned by
_MeaningBackwards = owns
Extension to
Stereotype: NestedConnectorEnd
Stereotype NestedConnectorEnd has Generalization relationship to SysML1.4::NestedConnectorEnd stereotype

NestedConnectorEnd is instaciated into toolbox as MyMDG::NestedConnectorEnd (UML::Nesting)=Containment.

I dont have any idea what i could check anymore, is there any obvious mistake?

[qwerty]

Try UCextend (lowet case E). Might be a clash with an existing stereotype. Guessing that since UCinclude uses a LC i.

q.

[vrieg]

no did not fix it. meanwhile i assume it is a bug.