Introducing C++

[aayush_005]

Some Remarks about Programming

Programming is a core activity in the process of performing tasks or solving problems with the aid of a computer. An idealised picture is:

[problem or task specification] - COMPUTER - [solution or completed task]

Unfortunately things are not (yet) that simple. In particular, the "specification" cannot be given to the computer using natural language. Moreover, it cannot (yet) just be a description of the problem or task, but has to contain information about how the problem is to be solved or the task is to be executed. Hence we need programming languages. Click here for a more detailed view of the problem solving pipeline.

There are many different programming languages, and many ways to classify them. For example, "high-level" programming languages are languages whose syntax is relatively close to natural language, whereas the syntax of "low-level" languages includes many technical references to the nuts and bolts (0's and 1's, etc.) of the computer. "Declarative" languages (as opposed to "imperative" or "procedural" languages) enable the programmer to minimise his or her account of how the computer is to solve a problem or produce a particular output. "Object-oriented languages" reflect a particular way of thinking about problems and tasks in terms of identifying and describing the behaviour of the relevant "objects". Smalltalk is an example of a pure object-oriented language. C++ includes facilities for object-oriented programming, as well as for more conventional procedural programming.


1.2 The Origins of C++


C++ was developed by Bjarne Stroustrup of AT&T Bell Laboratories in the early 1980's, and is based on the C language. The name is a pun - "++" is a syntactic construct used in C (to increment a variable), and C++ is intended as an incremental improvement of C. Most of C is a subset of C++, so that most C programs can be compiled (i.e. converted into a series of low-level instructions that the computer can execute directly) using a C++ compiler.

C is in many ways hard to categorise. Compared to assembly language it is high-level, but it nevertheless includes many low-level facilities to directly manipulate the computer's memory. It is therefore an excellent language for writing efficient "systems" programs. But for other types of programs, C code can be hard to understand, and C programs can therefore be particularly prone to certain types of error. The extra object-oriented facilities in C++ are partly included to overcome these shortcomings.


1.3 ANSI C++


The American National Standards Institution (ANSI) provides "official" and generally accepted standard definitions of many programming languages, including C and C++. Such standards are important. A program written only in ANSI C++ is guaranteed to run on any computer whose supporting software conforms to the ANSI standard. In other words, the standard guarantees that ANSI C++ programs are portable. In practice most versions of C++ include ANSI C++ as a core language, but also include extra machine-dependent features to allow smooth interaction with different computers' operating systems. These machine dependent features should be used sparingly. Moreover, when parts of a C++ program use non-ANSI components of the language, these should be clearly marked, and as far a possible separated from the rest of the program, so as to make modification of the program for different machines and operating systems as easy as possible.


1.4 The C++ Programming Environment in Linux

The best way to learn a programming language is to try writing programs and test them on a computer! To do this, we need several pieces of software:
An editor with which to write and modify the C++ program components or source code,
A compiler with which to convert the source code into machine instructions which can be executed by the computer directly,
A linking program with which to link the compiled program components with each other and with a selection of routines from existing libraries of computer code, in order to form the complete machine-executable object program,
A debugger to help diagnose problems, either in compiling programs in the first place, or if the object program runs but gives unintended results.

There are several editors available for Linux (and Unix systems in general). Two of the most popular editors are emacs and vi. For the compiler and linker, we will be using the GNU g++ compiler/linker, and for the debugger we will be using the GNU debugger gdb. For those that prefer an integrated development environment (IDE) that combines an editor, a compiler, a linking program and a debugger in a single programming environment (in a similar way to Microsoft Developer Studio under Windows NT), there are also IDEs available for Linux (e.g. V IDE, kdevelop etc.)

Appendix A.1 of these notes is a brief operational guide to emacs and g++ so that you can get going without delay. If you are interested in learning more about Linux you can click here.

(BACK TO COURSE CONTENTS)
1.5 An Example C++ Program

Here is an example of a complete C++ program:
// The C++ compiler ignores comments which start with
// double slashes like this, up to the end of the line.

/* Comments can also be written starting with a slash
followed by a star, and ending with a star followed by
a slash. As you can see, comments written in this way
can span more than one line. */

/* Programs should ALWAYS include plenty of comments! */

/* Author: Rob Miller and William Knottenbelt
Program last changed: 30th September 2001 */

/* This program prompts the user for the current year, the user's
current age, and another year. It then calculates the age
that the user was or will be in the second year entered. */

#include <iostream>

using namespace std;

int main()
{
int year_now, age_now, another_year, another_age;

cout << "Enter the current year then press RETURN.\n";
cin >> year_now;

cout << "Enter your current age in years.\n";
cin >> age_now;

cout << "Enter the year for which you wish to know your age.\n";
cin >> another_year;

another_age = another_year - (year_now - age_now);

if (another_age >= 0) {
cout << "Your age in " << another_year << ": ";
cout << another_age << "\n";
} else {
cout << "You weren't even born in ";
cout << another_year << "!\n";
}

return 0;
}

Program 1.5.1

This program illustrates several general features of all C++ programs. It begins (after the comment lines) with the statement

#include <iostream>


This statement is called an include directive. It tells the compiler and the linker that the program will need to be linked to a library of routines that handle input from the keyboard and output to the screen (specifically the cin and cout statements that appear later). The header file "iostream" contains basic information about this library. You will learn much more about libraries of code later in this course.

After the include directive is the line:

using namespace std;


This statement is called a using directive. The latest versions of the C++ standard divide names (e.g. cin and cout) into subcollections of names called namespaces. This particular using directive says the program will be using names that have a meaning defined for them in the std namespace (in this case the iostream header defines meanings for cout and cin in the std namespace).

Some C++ compilers do not yet support namespaces. In this case you can use the older form of the include directive (that does not require a using directive, and places all names in a single global namespace):

#include <iostream.h>

Much of the code you encounter in industry will probably be written using this older style for headers.

Because the program is short, it is easily packaged up into a single list of program statements and commands. After the include and using directives, the basic structure of the program is:
int main()
{
First statement;
...
...
Last statement;

return 0;
}

All C++ programs have this basic "top-level" structure. Notice that each statement in the body of the program ends with a semicolon. In a well-designed large program, many of these statements will include references or calls to sub-programs, listed after the main program or in a separate file. These sub-programs have roughly the same outline structure as the program here, but there is always exactly one such structure called main. Again, you will learn more about sub-programs later in the course.

When at the end of the main program, the line

return 0;

means "return the value 0 to the computer's operating system to signal that the program has completed successfully". More generally, return statements signal that the particular sub-program has finished, and return a value, along with the flow of control, to the program level above. More about this later.

Our example program uses four variables:

year_now, age_now, another_year and another_age

Program variables are not like variables in mathematics. They are more like symbolic names for "pockets of computer memory" which can be used to store different values at different times during the program execution. These variables are first introduced in our program in the variable declaration
int year_now, age_now, another_year, another_age;

which signals to the compiler that it should set aside enough memory to store four variables of type "int" (integer) during the rest of the program execution. Hence variables should always be declared before being used in a program. Indeed, it is considered good style and practice to declare all the variables to be used in a program or sub-program at the beginning.