Problem with assigning variables of type List

[Hans Nowak]

Greg Ewing wrote:
> Hans Nowak wrote:
> 
>>
>> OK. That's two ways of looking at the same thing. I prefer to think of 
>> things as objects with names bound to them, rather than pointers to 
>> objects in boxes.
> 
> But that only works when you *have* names. Many
> objects in a typical Python program don't have
> any names bound to them -- contents of lists,
> tuples, dicts, etc.
> 
> You need some concept of boxes with pointers
> in them to make sense of all those. Once you
> have that, it seems natural to me to think
> of variables as named boxes.

I'm growing a bit tired of the debate, and since there's obviously several ways 
to look at the same thing, it's kind of pointless to keep on argueing about 
what is "right" way. (This includes people who send me private mail too. :-) 
But here is one such way to look at it:

1. Everything is an object. How this is implemented in C (or Java, or Caml, or 
whatever the underlying language is), is largely irrelevant, unles you hack the 
C API (etc) of course.

2. Objects may have names bound to them. A name is simply a way to refer to the 
object, but an object can be retrieved in other ways too.

3. Code like

   x = "fubu"
   y = x

is very different from equivalent-looking Pascal or C code, whatever anyone 
might say. *Because* of this difference, I think the 
objects-possibly-with-names-bound-to-them approach makes sense. Learn this, and 
  everything will (should) be clear. Use the conventional approach, and it will 
be confusing, IMHO-- you'll have to resort to something like "they're still 
boxes just like in language X, they just have pointers in them".

Python doesn't have pointers, and I'm damn glad it doesn't. Let's not introduce 
them into an explanation of Python's internal model.

-- 
Hans (base64.decodestring('d3VybXlAZWFydGhsaW5rLm5ldA=='))
# decode for email address ;-)
The Pythonic Quarter:: http://www.awaretek.com/nowak/
Kaa:: http://www.angelfire.com/jazz/aquila/blog/blogger.html