Parent references: hot or not?

[yaipa h.]

Dylan,

Might be worth the look...

 Multiple dispatch  
  . Generalizing polymorphism with multimethods

 by David Mertz, Ph.D. (mertz at gnosis.cx)

 link,
    http://www-106.ibm.com/developerworks/linux/library/l-pydisp.html

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Dylan Reinhardt <python at dylanreinhardt.com> wrote in message news:<mailman.1050762692.5175.python-list at python.org>...
> Hi all,
> 
> I'm working on a project where I have multiple instances of a container
> class that each hold a large number of another class.
> 
> I need the "contained" object to be able to make use of methods and
> attributes in its container. Inheritance won't do the trick since the
> things I need to refer to will be pretty dynamic.
> 
> A very minimal example of what I'm doing:
> 
> ---------------
> 
> class Container:
>    def __init__(self):
>       self.items = {}
> 
>    def add_item(self, item_name):
>       self.items[item_name] = Item(self)
> 
>    some_attr = 'spam'
> 
> 
> class Item:
>    def __init__(self, parent):
>       self.parent = parent
> 
>    def some_method(self):
>       return self.parent.some_attr
> 
> ---------------
> 
> This seems to be working well, but I want to check my assumptions.
> 
> If I'm understanding correctly:
> 
> 1. Any Item's self.parent is only a *reference* to its parent.
> 2. Holding a reference to a parent isn't some kind of recursive paradox.
> 3. As I add Items, Container size will grow linearly, not exponentially
> 4. It will be impossible to fully delete a non-empty Container.
> 
> I think those are valid assumptions, but would appreciate any
> confirmation, contradiction, or insight anyone cares to offer.
> 
> Also, if there's some other consideration or reason why holding a
> reference to one's parent is a Bad Idea, I'd love to hear about it.
> 
> Thanks in advance,
> 
> Dylan