From: Richard Kilmer <rich@...>
Date: 2003-11-06T01:33:41+09:00
Subject: Re: Managing metadata about attribute types


On Nov 5, 2003, at 10:58 AM, Chad Fowler wrote:

> On Wed, 5 Nov 2003, Richard Kilmer wrote:
>
> #
> # This goes beyond just setting attributes.  Many methods need 
> parameters
> # that are not simply attr_accessor methods.  Those parameters are
> # going to be an object.  The question is whether you write your 
> methods
> # to expect a certain "namespace" (Class/Module), or just simply a set
> # of behaviors.
> #
>
> Thank you, Rich.  This is the first time I remember seeing classes and
> modules referred to as "namespaces".  In the Ruby world, this is really
> the way to think about them.  Classes and modules are, among other 
> things,
> a method of grouping behaviors.  And, like other namespaces, this set 
> of
> behaviors might change.  So, testing for the namespace is rarely 
> useful.

Right.  The big reason for testing the namespace for an object that you 
are
sending a message into is to disambiguate the message.  For example if 
you
had two classes:

class Tree
   def bark
     @bark
   end
end

class Dog
   def bark
     p 'Woof!'
   end
end

Sending the message "bark" on a Tree returns an attribute (the 'Bark'
instance) and the message "bark" on a Dog has the dog act ("Woof!").

Again, I find this ambiguity rarely creeps into a Ruby application, and
I have written many Ruby projects of size.