Abstraction + Encapsulation: What's the difference?

Abstraction and encapsulation are terms you often hear when learning object oriented programming. Both are about hiding details. But, what's the difference? What do they really mean?

Abstraction is a means of hiding details in order to simplify an interface.

So, using a car as an example, all of the controls in a car are abstractions. This allows you to operate a vehicle without understanding the underlying details of the steering, acceleration, or deceleration systems.

A good abstraction is one that standardizes an interface broadly, across multiple instances of a similar problem. A great abstraction can change an industry.

The modern steering wheel, brake pedal, and gas pedal are all examples of great abstractions. Car steering initially looked more like bicycle steering. And both brakes and throttles were operated by hand. But the abstractions we use today were so powerful, they swept the industry.

Encapsulation is a means of hiding details in order to protect them from outside manipulation.

Encapsulation is what prevents the driver from manipulating the way the car drives — from the stiffness of the steering, suspension, and braking, to the characteristics of the throttle, and transmission. Most cars do not provide interfaces for changing any of these things. This encapsulation ensures that the vehicle will operate as the manufacturer intended.

Some cars offer a small number of driving modes — like luxury, sport, and economy — which allow the driver to change several of these attributes together at once. By providing driving modes, the manufacturer is allowing the driver some control over the experience while preventing them from selecting a combination of attributes that would render the vehicle less enjoyable or unsafe. In this way, the manufacturer is hiding the details to prevent unsafe manipulations. This is encapsulation.

This article started it's life as an answer on StackOverflow.