The mechanical explanation is that the superclass knows if/when it is inherited by another class and can do something about it.
Ummm.. okay, but give me at least one example (even if that is not the full story) of why that allows you to do something otherwise not possible or why it is more elegant.
In the video we watched as part of this course, there was this example.
If I understand correctly, this allows an elegant way to set a class instance variable. Elegant in the sense that very little code exists at the subclass level (which is what the user will write, the superclass is framework).
The superclass not only contains the bulk of the code but also establishes a convention. In theory, each subclass can declare its own class instance variables, e.g.
@table_name = "volumes" but who would check if you made a typo? By having a method set_table_name, it is cleaner cuz typos will be caught at runtime and flagged.
# cheat_sheet_001.rb module ActiveRecord class Base def self.set_table_name(new_name) # This is a class inst var and not an inst var # because it is a class method (self.xxx) # When this method is called, we want self to be the subclass # and not Base class in order for this to work @table_name = new_name end def self.inherited(subcls) name = subcls.name.downcase + 's' # At this point, self is Base # Switch self to subclass by explicitly calling # set_table_name on subclass subcls.set_table_name(name) end end end class Book < ActiveRecord::Base set_table_name "volumes" end p Book.instance_variables #=> ["@table_name"] p Book.instance_variable_get(:@table_name) #=> "volumes"
Note: This tip was written by Himansu Desai. Thx!
This page was last updated on 20th Dec. 2009.