What is the use of combining data and functions in c plus plus?

Answer 1

In procedural programming, we define a set of functions that will act upon a given set of data. However, the data and the functions that affect that data are actually quite separate, which can leave the data "exposed"; functions are free to alter the data without any regard to the actual type of data or what it represents. Ownership of the data is ambiguous at best. Without safeguards, the data could be invalidated at any time.

In structured programming we can organise the data and procedures in a more systematic fashion. However the data still required safeguards to maintain its validity.

With object-oriented programming we can combine small sets of data with the methods that can specifically modify that data into a single entity; an object. Each object is solely responsible for its own data, which can now be safely hidden from procedures that have no business accessing that data. Interfaces to the data can be designed such that the underlying data remains in a valid state at all times and access to the data and the functions that can modify the data can be strictly controlled by the object itself.

Users of objects do not need to know how an object physically stores its data, nor how it manipulates that data -- their only concern is to store the data, retrieve the data, and to modify the data. All of this is achieved by the interface exposed by the object. They need not concern themselves with the underlying implementation of that object, only that it does what it was intended to do in a highly predictable manner, nothing more and nothing less.

As with structured programming, complex data structures can be created by combining existing objects to create new objects. But since every object takes care of its own data, highly complex and robust structures that would otherwise be extremely difficult to implement become feasible.

An object that contains other objects needn't be concerned with how those objects actually work. For instance, a list object is not concerned with the type of data it can store, its only concern is to manage the node objects in the list, nothing more and nothing less. Similarly, the individual node objects have no concern for the type of data they contain, only that they have data the user can retrieve. By delegating jobs to the objects that actually know how to implement the work, the risk of invalidating data is greatly reduced.