Supervenience as a Design Pattern

Single-inheritance object-oriented languages have been extended with features for adding methods to a class from other classes without disturbing the source code of the receiving unit. Well known examples include traits, mix-ins, default methods in interfaces, and method references. The purpose of these offerings is explained in terms of convenient re-use of individual instance methods. These are most applicable when inheritance of a needed method is not possible without altering the application’s class hierarchy. The intended objectives include: (a) avoid duplicating a method in another class, (b) modify a class or interface for another version of the system without forcing the first version to change, (c) allow dynamic composition of methods into a “class”, (d) accomplish each of these objectives without forcing the creation of a new class with a slightly different name or position in a namespace or hierarchy. These are certainly useful objectives in making practical coding choices. However, point solutions based on methods may lead class designs away from other principles supporting quality and maintainability of the design such as separation of concerns.