Association End - Target Scope - where is it from?

[Paolo F Cantoni]

I'm about to "Hijack" the Association End Target Scope property for my own (nefarious  ) purposes.

The drop-down has two values: instance and classifier.

I can't find where it (the concept) comes from - Google doesn't help. Anyone know how this property came about in the Sparx connector design?  Can anyone point me at a formal definition?

TIA,
Paolo

[Geert Bellekens]

Paolo,

In UML superstructure v 1.5 (I know, very old) There a attribute AssociatonEnd.TargetScope of type ScopeKind.
The specs say:

In the metamodel, ScopeKind defines an enumeration that denotes whether a feature
belongs to individual instances or an entire classifier. Its values are:
- instance:
      The feature pertains to Instances of a Classifier. For example, it is a distinct Attribute in each Instance or an Operation that works on an Instance.
- classifier:
      The feature pertains to an entire Classifier. For example, it is an Attribute shared by the entire Classifier or an Operation that works on the Classifier, such as a creation operation.

I haven't found that enumeration in UML 2.5, so my guess is that it is now obsolete (which is probably the answer you were hoping for )

Geert

[Paolo F Cantoni]

Paolo,

In UML superstructure v 1.5 (I know, very old) There a attribute AssociatonEnd.TargetScope of type ScopeKind.
The specs say:

In the metamodel, ScopeKind defines an enumeration that denotes whether a feature
belongs to individual instances or an entire classifier. Its values are:
- instance:
      The feature pertains to Instances of a Classifier. For example, it is a distinct Attribute in each Instance or an Operation that works on an Instance.
- classifier:
      The feature pertains to an entire Classifier. For example, it is an Attribute shared by the entire Classifier or an Operation that works on the Classifier, such as a creation operation.

I haven't found that enumeration in UML 2.5, so my guess is that it is now obsolete (which is probably the answer you were hoping for )

Geert

Hi Geert,
Yes, I found that also - but it relates to the feature (attribute) not the connector (although allowing that in v1.5 a "Named Association End IS an Attribute").  There doesn't seem to be such a property of the Attribute in EA v9+.
Anyway,  I've grabbed it for my purposes...  Although my use is, broadly, in line with the above

Paolo

[Eve]

Yes, I found that also - but it relates to the feature (attribute) not the connector (although allowing that in v1.5 a "Named Association End IS an Attribute").  There doesn't seem to be such a property of the Attribute in EA v9+.
Anyway,  I've grabbed it for my purposes...  Although my use is, broadly, in line with the above

Paolo

In UML 2 they are the same metaclass: "Property".

The enumeration mentioned doesn't exist in UML 2 because they replaced its use with a boolean property isStatic on the Feature metaclass.

[Paolo F Cantoni]

[SNIP]

In UML 2 they are the same metaclass: "Property".

The enumeration mentioned doesn't exist in UML 2 because they replaced its use with a boolean property isStatic on the Feature metaclass.

Yeah, that's what I figured, Simon - but I couldn't be sure.  Thanks for clarifying.  So, there's no problem with the hijack?  It seems it wasn't going anywhere soon...

Paolo

[Eve]

If you're asking if it's likely to stick around, I'm sure it will.

From a modeling perspective. As always, my recommendation would be to model information in the manner that will be the clearest to your intended audience. My concern with your proposed hijack is that it isn't visible on a diagram at all.

I would probably opt to use the stereotype field instead of hijacking another field, in part because it's visible on the diagram. Of course I wouldn't be surprised if you were already using the stereotype.

[Paolo F Cantoni]

If you're asking if it's likely to stick around, I'm sure it will.

From a modeling perspective. As always, my recommendation would be to model information in the manner that will be the clearest to your intended audience. My concern with your proposed hijack is that it isn't visible on a diagram at all.

I would probably opt to use the stereotype field instead of hijacking another field, in part because it's visible on the diagram. Of course I wouldn't be surprised if you were already using the stereotype.

I've REALLY moved away from overloading the Stereotype field.  However, the MAIN reasons I picked Target Scope is that it is available to the Shape Scripts.  THAT's how I surface the settings in Target Scope.

Paolo