Lecture - Imperative Programming TT23, II

• Loose coupling/High cohesion
  • Loose coupling: Different classes should not depend closely on the details of eachother
  • Principle of least knowledge: assume as little as you can about other objects
  • High cohesion: …
• Big idea: Polymorphism
  • Some constructs in programming languages should be able to process different objects with different types
  • Ad hoc polymorphism/overloading: define the same method for different types many different times
  • Parametric polymorphism/generics: code can be written generically so it can handle them generically without depending on their type
  • In OOP, having multiple components implement the same interface means that code can use different classes without worrying about each specific one
  • Useful for debugging, since you can mock components of a system
• Inheritance
  • Overriding methods in Scala works by putting override before the def
  • Then you use super.<orig-method> inside the method body to access the original
  • val a : <class> : new <subclass>. In this case, <class> is called the “static type” and <subclass> is called the “dynamic type”.
  • Can be dangerous when a subclass relies on the implementation details of a method in a the superclass (e.g. to call a method that they have overriden, to handle extra state). This is called the “fragile base class” problem.

Flashcards (GPT-4)