First-Class Relationships for Object-Orientated Programming Languages

The object-oriented paradigm describes a powerful systemwhere complex systems can be modelled by a network of communicating objects. The object model is powerful because it appeals to our understanding of the world, yet it has serious shortcomings where the model doesn’t match our expectations. In particular, relationships are completely ignored by most object-oriented programming languages. Programmers must rely on error prone code using references which are a poor substitute for many of the complex relationships programs must represent. There have been several attempts to add relationships to the objectoriented paradigm and object-oriented design and object-oriented design and specification languages now include relationships, however there has been little progress made in adding relationships at the implementation level. We believe this is because existing attempts have not refined the object-orientedmodel, but simply added relationships to existing systems. In this proposal we present a prototype for a novel model of relationships in the object-oriented paradigm in which relationships, rather than objects, are the dominant constructs. This shift in focus allows programmers to gain more leverage from the object-oriented abstraction by bringing the language closer to the systems which are modelled. In addition, our model produces a stronger hierarchy than existing object-oriented systems which makes program decomposition easier and reduces coupling between objects.