Data abstractions for data bases

Abstract

The concept of abstract data types has emerged from programming language research as a device to encourage and facilitate structured and modular programming [1]. It separates the abstraction of a data object from its implementation. The user of an abstract data type only concerns himself with the behavioral semantics of the type: what meaningful operations can be applied to objects of the type; the internal representation and structure of the type is unknown to him. In this way, irrelevent detail is suppressed and meaningfulness of programs enhanced. What application does this concept have to the area of data bases? There are some obvious similarities between the principle of abstraction and the concept of data independence of data bases [2]. Data independence means that application programs that utilize a data base do not need to know either the physical or the structural organization of the data base, but can relate to it purely on a logical plane. As with data abstractions, the goal is clarity and modifiability of programs, resulting from concentration on semantics rather than representation.