On 5/16/18 3:17 AM, Steven D'Aprano wrote:
On Wed, 16 May 2018 17:03:22 +1000, Ben Finney wrote:
So, no, I think the more useful – and less problematic – framing is that
every object *has* a value, and mutable objects may change to a
different value while remaining the same object.
What's an object?
That's not a rhetorical question. When I started learning Python, I had
*no idea* what "object oriented programming" was, or what classes and
objects and instances were. I still don't know if there is a precise,
accurate, unambiguous definition.
If you're going to describe Python in terms of objects for pedagogical
purposes, you better be prepared to explain in simple, plain English, non-
technical, unambiguous terms what an object is. Without reference to the
CPython (or any other Python) implementation.
It is true that "value" has a multitude of meanings. But "object" is even
worse. Not only does it also have a multitude of meanings (Websters lists
at least 7 for the noun, WordNet gives 5), but it ALSO has a technical
meaning in computer science, AND a specific meaning in Python as the name
of the type "object", AND also is used as a synonym for "instance".
I expect that most people understand that the ordinary English word
"value" has a multitude of meanings, and so will avoid thinking of it as
a precise technical term. I don't think you can say the same for "object".
I've also experimented with different ways to better say "everything is
an object". One possibility is, "any right-hand side of an assignment
is an object," though that is a bit tortured.
Now I'm thinking of trying, "Any piece of data is an object." I think
that might work well, with some complications arising when you have to
explain that functions (and modules and classes and ...) are also a kind
of data. At least it will let me use the word "data" so I can pretend
to be one of the cool kids :)
--Ned.
--
https://mail.python.org/mailman/listinfo/python-list
