Comment by layer8

The core of OOP is encapsulation (into objects) and polymorphic interfaces. You program against interfaces with well-defined contracts. Implementation details are encapsulated behind those interfaces. The same interface can have different implementations. The same interface-typed variable can point to different implementations at different times within the same program execution. The caller who invokes operations on an interface is supposed to not care about the implementation details behind it, and thus be decoupled from them. Interfaces can have an inheritance/subtyping relationship. (Implementations may as well, but that’s optional.) This enables abstracting over the commonalities of several related interfaces without having to introduce an adapter/proxy object. That’s basically it.