> I came up with some simple rules for assignment (variables as boxes with
> stuff in them), including a way of drawing them on the board, then had the
> kids 'execute' short snippets to see how they worked.  It's fun, it's
> active, and it doesn't cause monitor burn :)
> 
> Dustin

This all sounds like wholesome fun (there's that word again).

On the "variables as boxes with stuff in them" model, I think
we've touched on this earlier, but it might be useful to thread 
a bit on what're the most useful metaphors that'll not lead 
to confusion down the road.  I'm thinking of stuff like:

> >> a = b = [1,2,3,4]
> >> a
[1, 2, 3, 4]
> >> b
[1, 2, 3, 4]
> >> a[0]='A'
> >> a
['A', 2, 3, 4]
> >> b
['A', 2, 3, 4]

'a' and 'b' both point to the same object, and changing
the "contents" of 'a' also changes the contents of 'b'.
The "stuff in boxes" metaphor tends to "break" at this
point.

Another metaphor is "dog on a leash" and the idea that
more than one person has a leash to the same dog.  
Dog = Object in this picture.  So if you feed Joe's 
dog, you're maybe feeding Sally's dog as well, as 
they both have pointers to the same dog.

But other times, the "stuff in boxes" metaphor is apt
-- I'm not implying it has to be either/or, just that
we need a rich set of analogies handy, to keep Python
(and by extension, computer workings in general), as
transparent and intuitive as possible.  Dustin's point
that some mental model of computer memory is worth
instilling is right on I think.

I'm sure others have done some thinking along these 
lines and might elaborate on what they think is good
metaphorics.

Kirby