passing by refference

[Terry Reedy]

"Joshua Marshall" <joshway_without_spam at myway.com> wrote in message
news:ba0iqp$a0b$1 at ginger.mathworks.com...
> It's only here that I disagree with you.  I consider Python values
> themselves to be object references, which are simply copied/assigned
> to the formal parameters upon function invocation.

To me and, I believe, most pythoneers, the value of an int object is
its integer value: ..., -1, 0, 1, ... .  The value of a string object
is its sequence of bytes.  A 'reference' to a Python object is *not* a
Python object.  It is whatever a non-godly interpreter, human or
machine, needs to associate (non-object) names and container object
slots with Python objects.

> Where I do agree with you is that it's an implementation detail that
> these object references are implemented as C pointers.

Only in the C-coded interpreter.  The Java-coded interpreter does
something else.  No one seems to know quite how we humans associate
name with objects or if 'reference' will be part of some future
description.

> Ok, enjoyable conversation, but I think we're starting to retread
old
> ground.

So let me inject a somewhat different thought.  C, for example, is
pretty explicitly a language for computer execution: variables are
sized blocks of memory with integer addresses subject to addition and
subtraction.  (I am aware of the Intel exception, and of non-standard
kludges like 'near' and 'far'.)  Python is a more general algorithm
language defined in terms of names, typed information units (objects),
undefined, interpreter-implemented, 'associations' between the two,
and operations on the information values.  It is optimized more for
human interpretation than execution speed, but, unlike algorithm
pseudocodes, is also machine interpretable.  That's why I dubbed it
'executable pseudocode' over 6 years ago.   (The semi-oxymoronic irony
was intentional).  While some may not like this viewpoint, others like
me find it useful, which is what viewpoints are about.

Terry J. Reedy