Instance Variables Inheritance

Question

Can someone explain how a class can access the instance variables of its superclass and how that is not inheritance? I\'m talking about \'The Ruby Programming Language\' and the

Answer 1

You can compare these two examples.

example 1: it seems B inherit @i from A.

class A
  def initialize()
    @i = "ok"
  end
end

class B < A
  def print_i()
    p(@i)
  end
end

B.new().print_i()   # Shows "ok"

example 2: if B has its own initialize(), it cannot find @i.

class A
  def initialize()
    @i = "ok"
  end
end

class B < A
  def initialize()
  end

  def print_i()
    p(@i)
  end
end

B.new().print_i()   # nil, print nothing

class B's @i in example 1, is actually created implicitly by A#initialize() when you invoke B.new().

In your case, @x and @y in Point3D is actually created by Point#initialize(), not inherited from Point

Answer 2

It's confusingly worded. @x, @y, and @z are all instance variables on that Point3D instance. If that super(x,y) wasn't there, the Point3D instance would not have a @x or @y.

Answer 3

super(x,y) calls the constructor of the base class, which is the initialize method. If you take super(x,y) out, then the variables @x and @y won't appear in the derived class.

Answer 4

You are right, the book is wrong, or at least poorly worded

---

I would argue that the book is simply wrong, or at best, it's making a quite muddy explanation.

In all OO languages, the superclass and derived class don't have separate objects. When you create an instance of the derived class, it is also an instance of the superclass. There is one object and it is both classes at once.

Since there is only one object, there is only one set of instance variables.

This is the same as all other OO systems. The weird argument that book makes about how it just matters which method is run and how the methods themselves are what are really inherited does not add much in the way of clarity.

The problem with the terminology is that, sure, in a dynamically typed system there is no declaration in the first place, and so certainly the definition of the subclass doesn't inherit any field declarations ... because of course there aren't any. But just because there are no types to inherit doesn't make the opposite statement ("instance variables are not inherited") any more true, and it adds quite a bit of confusion because it implies that somehow the parent would have different instance variables, which is the nonsensical result of trying to talk about the objects the way they do.