Book a Demo

Author Topic: Association Classes on EAAggregations  (Read 3278 times)

Paolo F Cantoni

  • EA Guru
  • *****
  • Posts: 8626
  • Karma: +257/-129
  • Inconsistently correct systems DON'T EXIST!
    • View Profile
Association Classes on EAAggregations
« on: April 16, 2010, 04:08:29 pm »
In a (fairly) recent post: Aggregation and Association proposal, I discussed a way to resolve the ongoing issue of EA allowing both EAAggregations and UML Associations with AggregationKind.  I've heard nothing from Sparx, so I have to assume there's no progress on that front.

This post relates to the issue of AssociationClasses and EAAggregations.  Given that I can attach an Association Class to an Association with AggregationKind set, it would seem to be consistent to allow us to attach an AssociationClass to an EAAggregation.

I can see some issues for EA with this proposition.  EA uses the t_connector Subtype column (somewhat redundantly since - I think - it also uses the AggregationKind column) to denote the type of EAAggregation:  "Weak" (Shared) or "Strong" (Composite).  EA also uses the Subtype column for an Association to denote an AssociationClass (by setting the value to "Class").  Nevertheless, I feel this can be worked around without too much difficulty.

Since there seems to be no progress on solving the larger problem, the least that should happen is that EAAggregations and UML Association should be treated consistently.

Reported...
Thoughts?
Paolo
« Last Edit: April 16, 2010, 04:08:57 pm by PaoloFCantoni »
Inconsistently correct systems DON'T EXIST!
... Therefore, aim for consistency; in the expectation of achieving correctness....
-Semantica-
Helsinki Principle Rules!

Geert Bellekens

  • EA Guru
  • *****
  • Posts: 13495
  • Karma: +572/-33
  • Make EA work for YOU!
    • View Profile
    • Enterprise Architect Consultant and Value Added Reseller
Re: Association Classes on EAAggregations
« Reply #1 on: April 16, 2010, 04:33:42 pm »
Obvious, there's no reason why we shouldn't be able to add an associationclass to an EAAggregation.

Geert

Paolo F Cantoni

  • EA Guru
  • *****
  • Posts: 8626
  • Karma: +257/-129
  • Inconsistently correct systems DON'T EXIST!
    • View Profile
Re: Association Classes on EAAggregations
« Reply #2 on: April 16, 2010, 04:48:41 pm »
Quote
Obvious, there's no reason why we shouldn't be able to add an AssociationClass to an EAAggregation.

Geert
Yes, this came up because of my bug report: 854: Association Classes incorrectly imported.  I've got an long standing model (over 5 years old) which contains a mixture of Aggregations and Aggregated Associations.  I was trying to standardise and found I couldn't.

I'm going to have a look to see if I can do it via the DB...
[edit]As I predicted, EA gets "Multiple Personality Disorder" if you change the t_connector type from Association to Aggregation while it has the attached Associoation Class.  While the link disappears (after you save the dialog, the link in PDATA1 doesn't.[/edit]
Paolo
« Last Edit: April 16, 2010, 05:43:22 pm by PaoloFCantoni »
Inconsistently correct systems DON'T EXIST!
... Therefore, aim for consistency; in the expectation of achieving correctness....
-Semantica-
Helsinki Principle Rules!

Paolo F Cantoni

  • EA Guru
  • *****
  • Posts: 8626
  • Karma: +257/-129
  • Inconsistently correct systems DON'T EXIST!
    • View Profile
Re: Association Classes on EAAggregations
« Reply #3 on: April 16, 2010, 05:30:17 pm »
Interesting side-effect of the EAAggregation - Aggregated Association conflation, is that if you create an MDG Technology where some stereotypes extend Association and others extend Aggregation, the stereotype dropdown for both Associations and Aggregations contains BOTH sets of stereotypes!

In addition, since I use the _metatype feature when you open the properties the Association pretends to be an Aggregation (that is, the caption takes its value fro the stereotype (without reference to the metaclass).

Sparxians, you really need to get on top of this...  :-X

Paolo
Inconsistently correct systems DON'T EXIST!
... Therefore, aim for consistency; in the expectation of achieving correctness....
-Semantica-
Helsinki Principle Rules!