Sending a Message and Calling a Method
We often say that we call a method. Seen through the lens of object-oriented programming, a more precise description of what the sender does is send a message.
I learned this in Java, a statically typed language, where the two are easy to conflate. Comparing it with JavaScript helped me see the distinction.
0. A little terminology
// The return type, name, and parameters are part of the declaration. In discussing the interface, we’ll refer to the operation and its signature.
public void sayHello(String name) {
// The body supplies the implementation, also called the method.
System.out.println("안녕 " + name + "!");
}
1. Java checks types before the program runs
Java is statically typed. The compiler checks the type at compile time. When the sender tries to send a message, the receiver’s declared type has to expose the corresponding operation. Otherwise, the program never gets to send it. Compilation fails.
Working in a language like Java, it’s easy to come away thinking that sending a message and calling a method are the same thing.
2. JavaScript checks at runtime
JavaScript is dynamically typed. There’s no compile-time check that a sayHello() method exists on the receiver.
You can try sending the message to a rock object that has no such method. Ordinarily, trying to call it will then fail at runtime.
A receiver can even respond without defining the named method
It’s hard to explain the next example as a call to a method on rock. There’s no sayHello() implementation sitting on that object.
The message model makes it easier to follow. The sender sends a message named sayHello. The receiver intercepts it and decides how to respond at runtime.
References
Objects, Cho Young-ho, p. 49.
