From the book
From the book
From the book
PERSPECTIVE
In this section, we have described a fine and complete programming model that served (and still serves) many programmers for many decades. Modern programming, however, goes one step further. This next level is called data abstraction, sometimes known as object-oriented programming, and is the subject of the next section. Simply put, the idea behind data abstraction is to allow a program to define data types (sets of values and sets of operations on those values), not just static methods that operate on predefined data types.
Object-oriented programming has come into widespread use in recent decades, and data abstraction is central to modern program development. We embrace data abstraction in this book for three primary reasons:
- It enables us to expand our ability to reuse code through modular programming. For example, our sorts in Chapter 2 and binary search and other algorithms in Chapter 3 allow clients to make use of the same code for any type of data (not just integers), including one defined by the client.
- It provides a convenient mechanism for building so-called linked data structures that provide more flexibility than arrays and are the basis of efficient algorithms in many settings.
- It enables us to precisely define the algorithmic challenges that we face. For example, our union-find algorithms in Section 1.5, our priority-queue algorithms in Section 2.4, and our symbol-table algorithms in Chapter 3 are all oriented toward defining data structures that enable efficient implementations of a set of operations. This challenge aligns perfectly with data abstraction.
Despite all of these considerations, our focus remains on the study of algorithms. In this context, we proceed to consider next the essential features of object-oriented programming that are relevant to our mission.