What must a programmer know about an adt abstract data type to use it what values it can hold which operations it can perform how these operations are implemented?

Answer 1

Abstract data types (ADTs) are defined by base class designers. An ADT differs from an ordinary base class in that it has a pure-virtual interface. As with all base classes, the protected interface is intended for use by class implementers while the public interface is intended for all programmers (consumers).

All member data should be declared private. Even if there is no invariant associated with the data, declaring it private and providing a public interface ensures that any changes to the class representation will not affect consumers. Member data is an implementation detail that should be of no concern to either the class implementer or the consumer. Protected member data should always be treated with suspicion as this essentially exposes an implementation detail to implementers that could undermine encapsulation. Abstract data types in particular should have few data members, preferably none at all.

Class implementers are only concerned with the protected and public interfaces of their bass classes. Protected interfaces should be fully-documented as the implementer needs to know which methods need to be overridden and for what reason. A common convention is to place the protected interface default implementations in a separate translation unit, typically with a "_impl" suffix. This helps implementers examine the implementation details and thus eliminate redundant/duplicate code from their overrides. Nevertheless, a well-defined interface should be self-documenting and intuitive; the class header file should provide all the documentation necessary for an implementer or consumer to use the class without the need to examine the implementation details.