[Tutor] What is the difference between checking false?

[Dave Angel]

On 06/15/2013 11:53 PM, Jim Mooney wrote:
> On 15 June 2013 20:48, Joel Goldstick <joel.goldstick at gmail.com> wrote:
>
>> One and zero for True and False may seem not quite right today,
>
> I still think they should be taken out and shot ;')
>
> But my simplification plan failed. Equality always fails for different
> types, and 'not in front of None or any empty object such as empty
> string, empty list, empty tuple, or empty dict, is always True. But it
> appears that idea fails for an empty class:
>
> class NobodyHome: pass
> x = NobodyHome()
> print(not x) # Result is False when I thought this would be True.
>

The rules you're talking are not somehow mystically enforced by the 
compiler.  They're definitions and conventions that are presumably 
consistently applied in the library code for each builtin or library 
type.  But if you write your own, you get to make your own rules for if 
it's truthey or falsey.

Saying the class is empty is meaningless.  Even a trivial class has many 
methods that it inherits from object.  Anyway, if a class does not 
define its own truthy/falsey definition, then an instance of that class 
is always considered truthey.

The convention about list, tuple, dict, etc. is referring to 
collections, that implement a __len__() method.  Add such a method to 
your class, and you too can decide whether a particular instance is 
truthey or falsey.

But since your class isn't a collection, it'd be clearer to define the 
method __bool__(), which is what's tried first.  (In Python 3.x)

 >>> class NobodyHome:
...     def __bool__(self):
...         return False  #normally, you'd be testing some attribute to 
decide this

...
 >>> x = NobodyHome()
 >>> not x
True

-- 
DaveA