Activity Decomposition in MD SysML

Select an Activity and then use:

The result will differ depending on whether you use:

  1. explicit Pins (recommended for MD SysML)
  2. OR the "elided Pin" ObjectNode notation (discouraged)



Visit also:

Decomposition in MD SysML: Activity and Actions with Pins

Note use of Pin pairs of matching type. See next image for decomposition.

Decomposition of Activity HigherActivity (Pins owned as objects)

Note correspondence to in/out Pin pair, two properties are NOT owned by HigherActivity as an "ObjectNode".

MD SysML: decomposition of Activity with “ObjectNode”

Although the direct use of “ObjectNode” is discouraged in MD SysML, activity decomposition works in accordance with the SysML1.0 examples. See next image.

Decomposition of Activity HasObjectNode_AVOID_THIS

Corresponds with SysML1.0 examples, a single object appears to be owned by the usage context.

In fact a UML2 ObjectNode is abstract. Although the apparent ownership of an object by the higher activity context is consistent, it can't be represented using "elided Pin notation" in the activity diagram by MagicDraw UML.

Tuning ownership of properties in decomposed activities

By default the part properties generated by the The Activity Decomposition Wizard (for the object flows and actions owned by the decomposed Acfivity) are owned by the generated Associations, which Associations are owned by the same Namespace which owns the generated decomposition diagram.

The following diagrams show the default result and how to fine tune it. The first diagram shows the simple Activity to be decomposed. It includes two CallBehaviorActions and an (knowingly incorrect) "elided Pin" ObjectFlow, which is incorrectly represented by MagicDraw UML (and many other tools).


Make sure the general project option Enable dot notation for associations is on (true)
so you can see ownership of a Property on Association end by a Classifier.

Decomposition of Activity DecomposeMe (default result from Activity Decomposition Wizard)

Project settings (please note carefully):

  • The option show navigabilityis false for every Association end (all 6 of them)
    (inspection shows all 6 generated properties are non-navigable).
  • Global project option enable dot notation for Associations is true.


Only some Comments have been added to this diagram as generated by the wizard. Because project option enable dot notation for Associations is on we can tell that the generated part properties are NOT owned by the Classifier DecomposeMe (because there are no dots showing at the association ends).

Decomposition of Activity DecomposeMe: TUNED: ownership of the generated properties altered using navigability

Project settings (please note carefully): 

  • The global project option Change ownership of non-navigable association ends when changing navigability is set true.
  • The global project option Enable dot notations for associations is set true.
  • The option show navigability is set true for every Association end (all 6 of them).



On changing the navigability of the Association ends for Property o and b1 the ownership changes, they become owned by the activity DecomposeMe.

By contrast, the navigability of the Association end for Property b2 was left unchanged, it is still owned by the generated Association.