What are access specifiers in programming?

Answer 1

Access specifiers are used in object oriented programming to define public, protected and private interfaces. In non-object oriented languages, everything has public access. While this gives great freedom to the programmer, it makes it difficult for the programmer to maintain invariants. When distributing code, the onus is upon the consumer to observe those invariants and adhere to them. But the language will not assist them because invariants are typically established through user-defined comments which the compiler completely ignores.

Using access specifiers, the programmer can define an interface that ensures correct behaviour at all times. The onus of responsibility then shifts to the language itself. While this restricts consumer freedom, the restriction is intended to make objects easier to use without imposing any unnecessary limitations. The consumer is only granted access to what they actually need to access. Anything that could potentially violate an invariant is safely encapsulated within the object itself. If the consumer has access to the source code, they can easily examine the implementation should they wish to do so, but with a well-defined public interface, it is not necessary to know the private implementation details.

Access specifiers apply members of a class which includes base classes. Private members are only accessible to the class itself and to friends of the class (friendship can only be granted by the class itself). Protected members are the same as private members, but are also accessible to derivatives of the class. Public members are accessible to any code. Derivatives can also change the access specifiers of their bass class members, but they can only reduce access; they can never increase it.

In this way, class designers can ensure that objects of their class behave in a highly predictable manner. Consumers cannot break invariants encapsulated by a class, thus class implementations can be greatly simplified compared to public code, thus reducing the amount of code that needs to be generated. There is no need to constantly check an invariant holds because only the code that directly affects an invariant (constructors and mutators) also maintains that invariant. Class methods that simply grant access to object information (accessors) do not affect the invariant, so no error-checking is required; if the object exists, its invariant is assured to hold.