The data type created by the data abstraction process is called

Concept

Data abstraction is the design activity of describing a data type by what values it can hold and what operations may be performed on those values, while deliberately leaving out how the values are stored and how the operations are coded.

The product of that activity is therefore a specification — a set of values plus a set of operations, with the representation left unstated. A type defined in exactly this way is called an abstract data type (ADT): "abstract" because the specification is separated from every possible implementation, and "data type" because it still fixes the legal values and the legal operations.

Application

The stem asks for the name of what the abstraction process produces, not for the language construct later used to build it.

1. Take a concept such as a stack. Data abstraction states its values (a sequence of elements) and its operations (push, pop, top, isEmpty).

2. It also states the rules those operations must obey — here, that the element removed by pop is the one most recently added by push (LIFO order).

3. Nothing in that statement mentions an array, a linked list, a memory layout, or a single line of code. That omission is precisely the abstraction.

4. A type described as "values + operations + rules, representation unspecified" is named an abstract data type. So the data type created by the data abstraction process is called an abstract data type.

Cross-check

Contrast the result with the implementation-level and authorship-level terms:

• class and structure name concrete language constructs. A declaration of either commits to the stored fields — and, for a class, to the member functions that operate on them — which is exactly the information data abstraction leaves out. A class can be used to implement an ADT, and an abstract class can even publish an ADT's interface, but the construct itself sits at the implementation level; it is not what the abstraction process produces.

• user defined data type classifies a type by who declared it — the programmer rather than the language. A three-field record of integers is user-defined yet specifies no operations and hides nothing, so that label does not describe what abstraction produces.

Reading it the other way round confirms the same result: a stack, a queue, or a priority queue is standardly introduced as an abstract data type first, and only afterwards implemented with an array, a linked list, or a heap.