UML 2.4.1 RTF Avatar
  1. OMG Issue

UML241 — Can a Generalization really be in multiple GeneralizationSets?

  • Key: UML241-1
  • Legacy Issue Number: 15993
  • Status: closed  
  • Source: Model Driven Solutions ( Steve Cook)
  • Summary:

    According to the abstract syntax, Generalization::generalizationSet has upper bound *.

    According to the text:

    Package PowerTypes A generalization can be designated as being a member of a particular generalization set.

    There is only one place where the possibility of many sets is mentioned, where it says:

    “The generalizationSet association designates the collection of subsets to which the Generalization link belongs. All of the Generalization links that share a given general Classifier are divided into subsets (e.g., partitions or overlapping subset groups) using the generalizationSet association. Each collection of subsets represents an orthogonal dimension of specialization of the general Classifier.”

    The first of these sentences implies that a Generalization can belong to many (“collection of”) GeneralizationSets. The second sentence contains a phrase “subsets (e.g., partitions or overlapping subset groups)” that makes little sense. Rephrasing “subset groups” as “subsets” gives us “e.g., partitions or overlapping subsets” which seems to imply that the GeneralizationSets may overlap. But then “Each collection of subsets represents an orthogonal dimension of specialization” translates to “each collection of GeneralizationSets represents an orthogonal dimensionÂ…” which is obviously wrong. Rephrasing as “Each GeneralizationSet represents an orthogonal dimension Â…” makes more sense: but if they are orthogonal, how can they overlap?

    Then, in the notation and further explanations, there is no discussion whatsoever of the possibility of a generalization belonging to many GeneralizationSets.

    I think this is clearly an error in the metamodel.

  • Reported: UML 2.4 — Thu, 27 Jan 2011 05:00 GMT
  • Disposition: Resolved — UML 2.4.1
  • Disposition Summary:

    The text that the issue complains about is no longer in the spec. It’s clear that a Generalization can belong
    in more than one GeneralizationSet: “The generalizationSet property designates the GeneralizationSets to
    which the Generalization belongs.”
    However, there is a misleading phrase in the Notation that could be improved.

  • Updated: Fri, 6 Mar 2015 20:58 GMT