AggregationKind and end shapescripts

[Paolo F Cantoni]

Those of you who've been long-time members of the forum will recall my railing (over 5 years ago now) about how fundamentally broken the EAComposition and EAAggregation connectors were...
I've therefore religiously (certainly zealously) avoided them "like the plague" in the interim.  However, as I've previously indicated, I'm FORCED to use them - because there's no other way to achieve certain outcomes with MDG Technologies.

When you create an Association with shared or composite aggregation, you get the holonym diamond (indicating the AggregationKind) at the appropriate end.  Say the diamond is at the destination (target) end of the arc.  If you write a shape script which redefines the target shape, the AggregationKind indicator is NOT affected.  That's because it doesn't form part of the target shape.

This isn't true of the EAComposition and EAAggregation connectors.  The holonym diamond (indicating the AggregationKind) is the end shape!  Consequently, if you create a shape script for that end, you have to "manually" restore the holonym diamond yourself.

If you manually play with AggregationKind, EA responds in an interesting way...  You can set AggregationKind for either or both ends and EA will take the "Strongest" (where shared is a "Weak" Aggregation and composite is a "Strong" Aggregation) and apply that to one or other of the ends.  The end the diamond is applied to changes depending upon the settings.  Interestingly, if neither end has AggregationKind set, the holonym diamond is set to shared Aggregation.  So you can see that the potential for confusion is rife!

I think it would be good (read consistent) for EAAggregations and EAAssociations to behave the same way - end shape wise...

Thoughts?
Paolo

[Paolo F Cantoni]

See: Aggregation and Association proposal for a proposal to provide consistent processing and rendering of EA Aggregations, Compositions and Associations.

HTH,
Paolo