6.1 Two types of OOP system

6.1.1 Encapsulated OOP

• Python, C++, Java, RC & R6 in R…
• Methods belong to the objects or classes.
• Syntax to call a specific method: object.method(arg1, arg2, …).
• The class encapsulates both data(fields) and behaviors(methods).

6.1.2 Generic function OOP

• S3 & S4 in R
• Methods belong to generic functions.
• The class only encapsulates data - do not keep methods information.
• The generic function looks at the object’s class and behaves differently based on the class.
  • 一个特定的泛型函数，吃类A的对象，其执行的操作为A流程——相当于类A自己的方法；吃B的对象，其执行的操作为B流程——相当于类B自己的方法……
• Method dispatch is the process to find the correct method given a class.
  • In S3 and S4, polymorphic functions are used as generic functions.
  • Polymorphic functions behave differently for different input types

6.2 S3

6.2.1 Behind the generics

As we mentioned above, in S3, methods belong to function, which is called generic functions.

If a call to UseMethod() can be found in the source code of a function, then this function is a generic function. It is the function that figures out the correct method to call, the process of method dispatch.

>>> mean
function (x, ...)
UseMethod("mean")
<bytecode: 0x127f01628>
<environment: namespace:base>

What the generic function really does is to find and call a function named <generic>.<class> according to the class of the input. For example, Date method for the mean() generic is called mean.Date().

6.2.2 Testing generics