When to use nested classes and classes nested in modules?

Question

Other OOP languages have inner classes which cannot be instantiated without being bound to an upper level class. For instance, in Java,

class Car {
    class Wheel { }
}

only methods in the Car class can create Wheels.

Ruby doesn’t have that behaviour.

In Ruby,

class Car
  class Wheel
  end
end

differs from

class Car
end

class Wheel
end

only in the name of the class Wheel vs. Car::Wheel. This difference in name can make explicit to programmers that the Car::Wheel class can only represent a car wheel, as opposed to a general wheel. Nesting class definitions in Ruby is a matter of preference, but it serves a purpose in the sense that it more strongly enforces a contract between the two classes and in doing so conveys more information about them and their uses.

But to the Ruby interpreter, it’s only a difference in name.

As for your second observation, classes nested inside of modules are generally used to namespace the classes. For instance:

module ActiveRecord
  class Base
  end
end

differs from

module ActionMailer
  class Base
  end
end

Although this is not the only use of classes nested inside of modules, it is generally the most common.

In Ruby, defining a nested class is similar to defining a class in a module. It doesn't actually force an association between the classes, it just makes a namespace for the constants. (Class and Module names are constants.)

The accepted answer wasn't correct about anything. In the example below I create an instance of the lexically enclosed class without an instance of the enclosing class ever existing.

class A; class B; end; end
A::B.new

The advantages are the same as those for modules: encapsulation, grouping code used in only one place, and placing code closer to where it is used. A large project might have one outer module that occurs over and over in each source file and contains a lot of class definitions. When the various frameworks and library codes all do this, then they contribute only one name each to the top level, reducing the chance of conflicts. Prosaic, to be sure, but that's why they are used.

Using a class instead of a module to define the outer namespace might make sense in a one-file program or script, or if you already use the top level class for something, or if you are actually going to add code to link the classes together in true inner-class style. Ruby doesn't have inner classes but nothing stops you from creating about the same behavior in code. Referencing the outer objects from the inner ones will still require dotting in from the instance of the outer object but nesting the classes will suggest that this is what you might be doing. A carefully modularized program might always create the enclosing classes first, and they might reasonably be decomposed with nested or inner classes. You can't call new on a module.

You can use the general pattern even for scripts, where the namespace isn't terribly needed, just for fun and practice...

#!/usr/bin/env ruby

class A
  class Realwork_A
    ...
  end
  class Realwork_B
    ...
  end

  def run
    ...
  end
  
  self
end.new.run

You probably want to use this to group your classes into a module. Sort of a namespace thing.

for example the Twitter gem uses namespaces to achieve this:

Twitter::Client.new

Twitter::Search.new

So both Client and Search classes live under the Twitter module.

If you want to check the sources, the code for both classes can be found here and here.

Hope this helps!

There is yet another difference between nested classes and nested modules in Ruby prior to 2.5 that other answers failed to cover that I feel must be mentioned here. It is the lookup process.

In short: due to top level constant lookup in Ruby prior to 2.5, Ruby may end up looking for your nested class in the wrong place (in Object in particular) if you use nested classes.

In Ruby prior to 2.5:
Nested class structure: Suppose you have a class X, with nested class Y, or X::Y. And then you have a top level class named also Y. If X::Y is not loaded, then following happens when you call X::Y:

Having not found Y in X, Ruby will try to look it up in ancestors of X. Since X is a class and not a module, it has ancestors, among which are [Object, Kernel, BasicObject]. So, it tries to look for Y in Object, where it finds it successfully.

Yet it is the top level Y and not X::Y. You will get this warning:

warning: toplevel constant Y referenced by X::Y

Nested module structure: Suppose in the previous example X is a module and not a class.

A module only has itself as ancestor: X.ancestors would produce [X].

In this case, Ruby won't be able to look for Y in one of ancestors of X and will throw a NameError. Rails (or any other framework with autoloading) will try to load X::Y after that.

See this article for more information: https://blog.jetbrains.com/ruby/2017/03/why-you-should-not-use-a-class-as-a-namespace-in-rails-applications/

In Ruby 2.5:
Top level constant lookup removed.
You may use nested classes without fear of encountering this bug.

In the addition to previous answers: Module in Ruby is a class

$ irb
> module Some end
=> nil
> Some.class
=> Module
> Module.superclass
=> Object