7

I'm working through a book on Ruby, and the author used a slightly different form for writing a class initialization definition than he has in previous sections of the book. It looks like this:

class Ticket
  attr_accessor :venue, :date
  def initialize(venue, date)
    self.venue = venue
    self.date = date
  end
end

In previous sections of the book, it would've been defined like this:

class Ticket
  attr_accessor :venue, :date
  def initialize(venue, date)
    @venue = venue
    @date = date
  end
end

Is there any functional difference between using the setter method, as in the first example, vs. using the instance variable as in the second? They both seem to work. Even mixing them up works:

class Ticket
  attr_accessor :venue, :date
  def initialize(venue, date)
    @venue = venue
    self.date = date
  end
end
Greg Campbell
  • 15,182
  • 3
  • 44
  • 45
michaelmichael
  • 13,755
  • 7
  • 54
  • 60

1 Answers1

11

Since the setter method has been defined by attr_accessor and thus does nothing but setting the variable, there is no difference between using the setter method and setting the variable directly.

The only upside to using the setter method is that if you should later change the setter method to do something more than setting the variable (like validate the input or log something), your initialize method would be affected by these changes without you having to change it.

sepp2k
  • 363,768
  • 54
  • 674
  • 675
  • +1 I've found that change notification isn't possible (or is difficult and undocumented) with `@venue`, only with `self.venue`. If anyone knows of a way to observe attribute-based setters, let me know ;) – Joseph Yaduvanshi May 02 '10 at 02:24