UML data models from an ORM perspective : Part 10

This paper is the tenth in a series of articles examining data modeling in the Unified Modeling Language (UML) from the perspective of Object Role Modeling (ORM). Part 1 discussed historical background, language design criteria, object reference and singlevalued attributes. Part 2 covered multi-valued attributes, basic constraints, and instantiation using UML object diagrams or ORM fact tables. Part 3 compared UML associations and related multiplicity constraints with ORM relationship types and related uniqueness, mandatory role and frequency constraints, as well as how associations may be instantiated. Part 4 contrasted ORM nesting, co-referencing and exclusion constraints with UML association classes, qualified associations, and xor-constraints respectively. Part 5 discussed subset and equality constraints. Part 6 discussed subtyping. Part 7 discussed value, ring and join constraints. Part 8 listed some recent updates to the UML standard, then discussed aggregation. Part 9 examined initial values and derived data in ORM and UML. Part 10 discusses changeability and collection types in UML and ORM.