Closures Are Kind of Like Objects
What Do People Mean By An Object?
Lots of things!
- A collection of data
- A grouping of data and functions that act on that data
- When many objects have the same structure of data and/or functions we might start thinking of them as members of a class/type
- When objects have their data accessed by their functions we start talking about "encapsulation".
- When object types can share their structures we start talking about inheritance/etc.
So Objects can be typed or untyped.
Let's Look At A Closure!
Compare what is gained and lost.
(define (adder x)
(lambda (y) (+ x y)))
class PythonClass(object):
__slots__ = ('x')
def __init__(self, x):
self.x = x
def add_x_to(self, y):
return self.x + y
Comparison
-
We have state, x
- We have encapsulation
-
We lost a name
- One is a function with some state, the other is some named state and a named function
When To Use?
- When an API expects a function (ex: a callback)
- When you need to preserve state, but only one time, a closure is a lighter weight solution than an entire class.
- When you want much stricter enforcement of encapsulation.