Object-Relational Programming

Rationale

While the relational data model and its commercial implementations have undergone considerable extensions to become object-relational, comparable advances from mainstream object-oriented programming languages seem to be rare. Complementing recent efforts to integrate SQL-like querying of memory-based collections into languages such as C# and Java, we specify a gentle extension of the object-oriented programming model with bidirectional relationships whose use in a program is independent of their multiplicities.

Breakdown

Because of the extent of the undertaking, it has been broken down into different parts, or phases.

• Phase 1: Definition and implementation of to-many pointers for replacing collections in the implementation of relationships to many objects in OOP.
• Phase 2: Attachment of to-many pointers to roles as special traits from which relationships and associated behaviour are inherited.
• Phase 3: Development of pairs of roles (role and counterrole) to relationships as a new, first-class programming construct.

Host Language

Because of the dependence on traits as containers of relationships and associated behaviour, our host language is Scala.