Control Structures ------------------ Up to now, all of the programs we have seen were examples of what is known as STRAIGHTLINE programming. Program execution commenced with the first statement and proceeded statement by statement until the end of the program was reached. Most non-trivial programs require some sort of mechanism to alter the sequence of statement execution. C++ has two conditional constructs and three looping constructs. The conditional constructs determine which statements are to be executed depending upon a certain condition, the looping contructs determine the number of times a block of statements is to be executed. The Boolean type --------------- In the original version of the C programming language a LOGICAL FALSE was represented by the integer value 0, a LOGICAL TRUE was represented by any NON-ZERO integer. Thus in C (and early versions of C++) the following were all legal ways of representing a logical true: 1 5 678*3 Similarly, the following are legal ways of expressing a logical FALSE 0 16 % 16 C++ now has the built-in BOOLEAN logical data type which is named bool. The bool data type can take on one of two possible symbolic constants: true, false. Relational Operators -------------------- C++ has two type of relational operators: equality and ordering. The equality operators are: == returns true if both of the operands are equal to each other != returns true if the two operands are not equal to each other Some examples of the use of the equality operators are: int i = 1; int j = 1; int k = 2; (i==j) ; // this will return true (i==k) ; // this will return false (i!=k) ; // this will return true We are not limited to comparing only integers for equality or inequality. Any built-in C++ data type can be used (when working with floating point values we should be a bit more careful... more on this later). char letter1, letter2, letter3; letter1 = 'a'; letter2 = 'b'; letter3 = 'c'; letter1 == letter2 ; // this will return false; letter2 != letter3 ; // this will return true; What do you think the output of the following program which tests two structs for equality will be ? // Author: Ted Obuchowicz // Jan. 16, 2002 // example program illustrating use of // equality operators on struct data type #include #include using namespace std; int main() { struct a_simple_struct { int a_number; char a_letter; bool a_boolean; }; a_simple_struct first_struct, second_struct, third_struct; first_struct.a_number = 1; first_struct.a_letter = 'a'; first_struct.a_boolean = true; second_struct.a_number = 1; second_struct.a_letter = 'a'; second_struct.a_boolean = true; third_struct.a_number = 5; third_struct.a_letter = 'a'; third_struct.a_boolean = true; if (first_struct == second_struct) cout << "first and second structs are equal" << endl; else cout << "first and second structs are not equal" << endl; return 0; } The answer is nothing! The program DOES NOT EVEN COMPILE : ted@flash Programs 8:25pm >g++ -o struct_equality struct_equality.C struct_equality.C: In function `int main()': struct_equality.C:40: no match for `main()::a_simple_struct & == main()::a_simple_struct &' The compiler is complaining about line 40 in the source code which is the if statement: if (first_struct == second_struct) If you want to compare structs, then the comparison must be done on a field-by field basis. For example, the above program can be fixed by simply replacing the if statement with something along the lines of: if ((first_struct.a_number == second_struct.a_number) && (first_struct.a_letter == second_struct.a_letter) && (first_struct.a_boolean == second_struct.a_boolean) ) cout << "first and second structs are equal" << endl; else cout << "first and second structs are not equal" << endl; Ordering Operators ------------------ C++ uses the following 4 ordering operators to determine the relative size of two operands: < less than > greater than <= less than or equal >= greater than or equal Examples: ( 1 < 2 ) will return true ( 1 > 2) will return false Boolean Operators ----------------- C++ has three Boolean operators: &&, ||, and !. Their meaning is as follows: && : used to perform a logical AND of two boolean expressions: ex. (i < 3 ) && (k < 56 ) || : used to perform a logical OR of two expressions: ex. (i < 2 ) || ( k > 45) ! : used to perform a logical NEGATION ex. ! (i < 2) SHORT-CIRCUIT EVALUATION ------------------------ When evaluation the value of a logical expression, C++ requires that first the left operand be evaluated to yield a boolean value, then the right operand be evaluated. Furthermore, if the value of the entire logical expression may be determined from the left operand, then the right hand operand should not be evaluated. This is known as short-circuit evaluation. For example: ( i != 0) && ( (k/i) < 2 ) the ( i != 0) will be equal to false only in the case of i being equal to 0, since the boolean && operator requires both its operands to be true in order for the entire to be expression to be true, it is not even necessary to evaluate the right-hand side ( (k/i) < 2 ) in the case of i being equal to 0. By virtue of short-circuit evaluation, we can guarantee that division by zero will NEVER occur. If C++ did not provide short-circuit evaluation, then there would be a possibility that a division by zero would be attempted while evaluation the right hand side operand. Comparing floating point numbers ------------------------------- Round-off errors due to inherent limitiations of how real numbers are internally stored in a computer's memory can cause some problems when comparing floating point numbers for equality. For example, consider the following program which adds 0.0000001 ten times to theoretically give the value 0.000001: // Author: Ted Obuchowicz // Jan. 16, 2002 // example program illustrating use of // testing real numbers for equality #include #include #include using namespace std; int main() { float number; number = 0.0000001 + 0.0000001 + 0.0000001 + 0.0000001 + 0.0000001 + 0.0000001 + 0.0000001 + 0.0000001 + 0.0000001 + 0.0000001; if (number == 0.000001 ) cout << " number is equal to 0.000001" << endl; else cout << " number is not equal to 0.000001" << endl; cout << setprecision(15) << "number is equal to " << number << endl; return 0; } The output may surprise you: number is not equal to 0.000001 number is equal to 9.99999997475243e-07 When testing real number for equality, it is a better practice to check that the two numbers are close enough to one another. The following program makes use of the fabs() function found in the

library to check that the
absolute value of number - 0.000001 is less than some prespecified value called
epsilon:

// Author: Ted Obuchowicz
// Jan. 16, 2002
// example program illustrating use of
// testing real numbers for equality


#include #include #include #include using namespace std;



int main()
{

float number;
const float epsilon = 0.001;

number = 0.0000001 + 0.0000001 + 0.0000001 + 0.0000001 
      + 0.0000001 + 0.0000001 + 0.0000001 + 0.0000001 + 0.0000001 + 0.0000001;

if  (  fabs( (number - 0.000001) <= epsilon  )  )
   cout << " number is equal to 0.000001" << endl;
else
   cout << " number is not equal to 0.000001" << endl;


cout <<  setprecision(15)   << "number is equal to " << number << endl;

return 0;
}

We use the #include statment to tell the compiler where to look for
the definition of the fabs function. We would compile this program on the
ECE UNIX  g++ compiler with the following command line:

g++ -o floating_point_equality2 floating_point_equality2.C -lm

The -lm option must appear at the end of the command line. It tells the
linker to link the math library with your object code.

This time the program output is what we expect:

 number is equal to 0.000001
number is equal to 9.99999997475243e-07



The IF STATEMENT
---------------

The previous programs have already examined the use of the
IF statement.  The if statement comes in two forms:

Form 1:  if (expression)
             statement;

Form 2:  if (expression)
             statement;
         else
             another statement;

NOTE:  If it is required to perform more than one statement in a branch of
       the if statement, then the entire block of statements must be enclosed
       in { } such as:

       if (expression)
          {
            statement1;
            statement2;
            ...
            statementn;
           };
 
       or using Form 2:

        if (expression)
          {
            statement1;
            statement2;
            ...
            statementn;
          };
        else
          {
            statementx;
            statementy;
            ...
            statementz;
          };






Be careful to use the delimiting { } properly. Failure to
do so may cause strange program behaviour as is the 
following program:

// Author: Ted Obuchowicz
// Jan. 16, 2002
// example program illustrating INCORRECT use of
// if statement



#include #include using namespace std;



int main()
{

int i ;

cin >> i;


if (i != 1)
     cout << " i is not equal to 1" << endl;
     cout << " end of program goodbye" << endl;



return 0;
}


When we run this program and enter 1 for the value of 1,
the output is:

 end of program goodbye


What went wrong? The expression ( i != 1) is false for i equal to 1
(since 1 is equal to 1). So why was the "end of program goodbye"
produced as output? .  What the program actually did was:

if (i != 1)
     cout << " i is not equal to 1" << endl;


cout << " end of program goodbye" << endl;


The second cout is actually independent of the first if since we did
not delimit the two statements within { }.

If we change the program to:

// Author: Ted Obuchowicz
// Jan. 16, 2002
// example program illustrating INCORRECT use of
// if statement



#include #include using namespace std;



int main()
{

int i ;

cin >> i;


if (i != 1)
    {
     cout << " i is not equal to 1" << endl;
     cout << " end of program goodbye" << endl;
    };  // the above messages will be printed ONLY FOR values of i not equal to 1
        // when i is equal to 1, no output will be produced

     

return 0;
}


A sample session with the running program is:

ted@flash Programs 10:08pm >if_statement2
1
ted@flash Programs 10:09pm >


ted@flash Programs 10:09pm >if_statement2
3456
 i is not equal to 1
 end of program goodbye
ted@flash Programs 10:09pm >





CAVEAT
-------

The following is a VERY COMMON MISTAKE THAT BEGINNER C and C++
programmers make:


// Author: Ted Obuchowicz
// Jan. 16, 2002
// example program illustrating use of
// common programming error with an if 



#include #include using namespace std;



int main()
{

int balance_owed;

cin >> balance_owed;

if (balance_owed = 0)
  cout << "You owe me nothing" << endl;
else
  cout <<  "You owe me " << balance_owed << " dollars" << endl;


return 0;
}


This program when run will ALWAYS print out:

You owe me 0 dollars

irrespective of the value the user inputs for the integer variable
balance_owed!

WHY????

look very carefully at the if statement:

if (balance_owed = 0)
    ^^^^^^^^^^^^^^^^^

  cout << "You owe me nothing" << endl;
else
  cout <<  "You owe me " << balance_owed << " dollars" << endl;


the (balance_owed = 0) is an ASSIGNMENT of 0 to the variable balance_owed.
The result of the assignment is that variable balance_owed take son the value
0. The condition then reduces to:

if(0)
   cout << "You owe me nothing" << endl;
else
   cout <<  "You owe me " << balance_owed << " dollars" << endl;

since a 0 is false in C++, the condition is always false so the
else part is always executed with a value of 0 for the amount owed!!!

/********************************************************************/
/*                                                                  */
/*  ALWAYS USE == FOR COMPARISON IN IF STATEMENTS!!!!!              */
/*                                                                  */
/*                                                                  */
/********************************************************************/

The proper version of the program is:


// Author: Ted Obuchowicz
// Jan. 16, 2002
// example program illustrating use of
// correct use of if statement





#include #include using namespace std;



int main()
{

int balance_owed;

cin >> balance_owed;

if (balance_owed == 0)
  cout << "You owe me nothing" << endl;
else
  cout <<  "You owe me " << balance_owed << " dollars" << endl;


return 0;
}



Note that the only difference between the good and the bad versions
is the difference between = and == .  Amazing what a single missing
= will do in a program! PAY YOUR ATTENTION INSIDE OF IF STATEMENTS
AND ALWAYS CHECK FOR EQUALITY WITH THE ==   !!!!!!!

The proper output is:

ted@flash Programs 10:47pm >good_if
0
You owe me nothing
ted@flash Programs 10:47pm >good_if
456
You owe me 456 dollars



Nested if-else-if statements:
----------------------------

Consider the following if construct:

char command;
cin >> command;

if (command == 'u')
    cout << "Move up command was received" << endl;
else
   if (command == 'd')
      cout << "Move down command was received" << endl;
   else
      if (command == 'l')
         cout << "Move left command was received" << endl;
      else
         if (command == 'r')
            cout << "Move right command was received" << endl;
         else
            cout << "Invalid command" << endl;   

The indentation of the if statements is purely cosmetic. The compiler does not care about
nesting. We can rewrite the above if statement as:

if (command == 'u')
        cout << "Move up command was received" << endl;
else  if (command == 'd')
        cout << "Move down command was received" << endl;   
else  if (command == 'l') 
        cout << "Move left command was received" << endl;
else  if (command == 'r')
         cout << "Move right command was received" << endl;
else   
         cout << "Invalid command" << endl; 

Using this style improves program readability and avoids the problem of running out
of space on the right hand side as the nesting level of the if statements increases.


The switch statement
-------------------

C++ has a very convenient statement for multi-branch conditional testing as in the
above statement. It is called the switch statement.  The above  if can be rewritten
as a switch statement.

// Author: Ted Obuchowicz
// Jan. 17, 2002
// example program illustrating use of
//  switch statement



#include #include using namespace std;



int main()
{

char command;
cin >> command;

switch (command)
{
 case 'u':   cout << "Move up command was received" << endl;
             break;
 case 'd':   cout << "Move down command was received" << endl;    
             break;
 case 'l':   cout << "Move left command was received" << endl;
             break;
 case 'r':   cout << "Move right command was received" << endl;
             break;
 default :   cout << "Invalid command" << endl;
}

return 0;

}


When the switch statment is executed, the value of command is compared with
the specified cases. If a match is found, the statements corresponding to the
matching case are executed.  If the value of command does not match any of the
specified cases, then the sequience of statements for the specified default case
is executed (if there is no default case specified, flow of control resumes with the
statement immediately following the switch).

The break statement
------------------

Note that in the above switch, there is a break statement associated with every
specified case.  The break statement causes a change of flow of control to the
statement following the switch.  The break can be interpreted as "the switch statement
has completed its task and program execution should resume with the statement immediately
following the switch statement".

If a switch statement does not contain breaks for the specified cases, then the
action for the FIRST matching value is executed, FOLLOWED BY ANY REMAINING groups
of action statements (for the remaining values even though they do not match).

For example,

// Author: Ted Obuchowicz
// Jan. 17, 2002
// example program illustrating use of
//  switch statement



#include #include using namespace std;



int main()
{

char command;
cin >> command;

switch (command)
{
 case 'u':   cout << "Move up command was received" << endl;
 case 'd':   cout << "Move down command was received" << endl;    
 case 'l':   cout << "Move left command was received" << endl;
 case 'r':   cout << "Move right command was received" << endl;
 default :   cout << "Invalid command" << endl;
}

return 0;
return 0;

}


If the user supplies u for the value of the command, the program produces the
following output:

ted@flash Programs 12:55pm >switch_no_break
u
Move up command was received
Move down command was received
Move left command was received
Move right command was received
Invalid command


If we supply d for the value of the input variable command, the following output
is produced:

ted@brownsugar Programs 7:21pm >switch_no_break
d
Move down command was received
Move left command was received
Move right command was received
Invalid command


giving l as an input value results in:

ted@brownsugar Programs 7:21pm >switch_no_break
l
Move left command was received
Move right command was received
Invalid command


if we give r as the value :

ted@brownsugar Programs 7:25pm >switch_no_break
r
Move right command was received
Invalid command



Cases in a switch with no break
------------------------------

In most programming problems, every 'action' statement should 
have an associated break statement  as in

switch (condition)
{
 case value1 : action1;
               break;
 case value2 : action2;
               break;
 ...
  
 case valuen : actionn;
               break;

 default : default_action;
}

In some programming situations, using the so-called "fall through"
behaviour (some actions without an associated break) makes sense.
Consider once again the menu switch where we now consider the
upper and lower case letters as being valid input characters
(i.e. both 'u' and 'U' should be recognized as a legal move up
command, 'd' and 'D' as legal move down commands etc).  We can rewrite
our  previous program as:



#include #include using namespace std;



int main()
{

char command;
cin >> command;

switch (command)
{
 case 'u':  // fall through to the 'U' case
 case 'U':   cout << "Move up command was received" << endl;
             break;
 case 'd':  // fall through to the 'D' case
 case 'D':   cout << "Move down command was received" << endl;    
             break;

 case 'l':  // fall through to the 'L' case
 case 'L':   cout << "Move left command was received" << endl;
             break;
 case 'r':  // fall through to the 'R' case
 case 'R':   cout << "Move right command was received" << endl;
             break;
 default :   cout << "Invalid command" << endl;
}

return 0;

}


It should be noted that the switch statement in the above example
can be rewritten more clearly as:

case 'u' : case 'U' : cout << "Move up command was received" << endl;
                      break;

case 'd' : case 'D':  cout << "Move down command was received" << endl;    
                      break;

case 'l':  case 'L':  cout << "Move left command was received" << endl;
                      break;
case 'r':  case 'R':  cout << "Move right command was received" << endl;
                      break;
default :   cout << "Invalid command" << endl;
}


Using this style clearly indicates the code which is to be executed
for common cases.  


To reiterate, the switch statement is different from similar statements in 
languages such as Pascal in a very important way. In C and C++, when a program 
jumps to a certain line in a switch statement (when it finds the first matching
value of the listed cases) , the program THEN SEQUENTIALLY EXECUTES all the
statements following  that line unless you EXPLICITLY TELL OTHERWISE 
(with the break) .  Program execution DOES NOT AUTOMATICALLY stop at the next
case.  To make execution stop at a particular line, you must use the break
statement.  The break causes the program to continue exection from the next
statement following the switch.

One final note regarding the switch statement.  Both the expression which
is being tested and the values of the cases should be of an INTEGRAL type.
This limitation means that a switch statement cannot be used to determine
the flow of program control based on the value of a floating-point variable.


LOOPING CONSTRUCTS
-----------------

Suppose we wish to find the sum , and average value of a list
of  4 numbers.  Our first (brute-force) program might look like:

// Author: Ted Obuchowicz
// Jan. 21, 2002
// example program illustrating use of
// brute force approach to finding the sum and average of 5 floating
// point values



#include #include using namespace std;



int main()
{

float sum;
float average;
float number_1;
float number_2;
float number_3;
float number_4;
float number_5;

 
cout << "Please enter first number ";
cin >> number_1;

cout << "Please enter second number ";
cin >> number_2;

cout << "Please enter third number ";
cin >> number_3;

cout << "Please enter fourth number ";
cin >> number_4;

cout << "Please enter fifth number ";
cin >> number_5;

sum = number_1 + number_2 + number_3 + number_4 + number_5;

average = sum / 5.0;

cout << "The sum of the 5 numbers is : " << sum << endl;
cout << "The average value is : " << average << endl;

return 0;

}



Suppose we had to find the sum and average of a list of 10 000 values?
Clearly, the above brute force approach is not very convenient to use.

We can use a while loop to control the number of times a portion of
program code is to be executed.  The following is an ALGORITHM which
repeatedly gets the next input value and adds it to the sum of the
previous input values and then calculates the average:


Step 1:  Set the value of the running sum to 0.0

Step 2:  Set the value of the number of processed values to 0.

Step 3:  while (number of processed items < sixe of the list )
            Step 3a:   read in a value;
            Step 3b:   sum = sum + value;
            Step 3c;   add 1 to the number of items processed so far

Step 4:   calculate the average as sum / list size

Step 5:  display the values of the sum and the average


A complete C++ program which corresponds to the above pseudocode is:

// Author: Ted Obuchowicz
// Jan. 21, 2002
// example program illustrating use of
// while loop to find sum and average of 5 floating point values read in
// from keyboard



#include #include using namespace std;



int main()
{

float sum;
float average;

const int size_of_list= 5;
float number;
int number_of_items_processed;

// perform some initialization

sum = 0.0;
number_of_items_processed = 0;

while (number_of_items_processed < size_of_list)
{
  cout << "Please enter a number " << endl;
  cin >> number;
  sum = sum + number;  // update running sum
  number_of_items_processed++; // increment number of items processed so far
}

average = sum / size_of_list;

cout << "The sum is " << sum << endl;
cout << "The average is "  << average << endl;


return 0;
}


Improving the above program:

The above program is fine for lists whose sizes is known ahead of time.
What if we wanted to find the sum and average of an unknown number of
values????

We can modify the above program to test the number which is input to
see if it is a special "flag" ( in computer programming a flag variable
is one  which the programmer places "special meaning" to).  For example,
we can reserve the number -9999 to indicate that the list has no more values.
The modified program is :

// Author: Ted Obuchowicz
// Jan. 21, 2002
// example program illustrating use of
// while loop to find sum and average of 5 floating point values read in
// from keyboard



#include #include using namespace std;



int main()
{

float sum;
float average;

const float flag = -9999.0;
float number;
int number_of_items_processed;

// perform some initialization

sum = 0.0;
number_of_items_processed = 0;

cout << "Please enter a number " << endl;
cin >> number;

while ( ! ( (number - flag) < 0.00001 ) )
{
  
  sum = sum + number;  // update running sum
  number_of_items_processed++; // increment number of items processed so far
  cout << "Please enter a number " << endl;
  cin >> number;
  
}



if (number_of_items_processed != 0)
{
  average = sum / number_of_items_processed;
  cout << "The sum is " << sum << endl;
  cout << "The average is "  << average << endl;
}
else
{
  cout << " No numbers were entered " << endl;
  cout << " Sum is " << sum << endl;
  cout << " Impossible to compute average... sorry " << endl;
}



return 0;
}


Some comments on the above program:
----------------------------------

1) Note that before we enter the while loop, we first read in a
   number... this is necessary so that the variable called number
   has an intial value when it is being tested inside the conditional
   expression of the while loop:

cout << "Please enter a number " << endl;
cin >> number;

while ( ! ( (number - flag) < 0.00001 ) )
             -----
               |
                -----> this must have some value before it can be tested

2) the test in the while loop does not test for equality with -9999.0,
   rather it tests that the difference between the number read in and
   the constant flag of -9999 is less than some small number 
   ( actually we should really test for the absolute value of the
     difference between number and flag being less than some value...)

3) in the while loop , we read in a new value for number:

   cin >> number;

    this is necessary , since  we eventually want the while loop to
   terminate.

4) we test that some numbers have been actually entered to avoid a potential
   division by zero in our source code:

   if (number_of_items_processed != 0)
{
  average = sum / number_of_items_processed;
  cout << "The sum is " << sum << endl;
  cout << "The average is "  << average << endl;
}
else
{
  cout << " No numbers were entered " << endl;
  cout << " Sum is " << sum << endl;
  cout << " Impossible to compute average... sorry " << endl;
}




Alternative solution
-------------------

While flags are useful programming tricks, they are not totally
without problems of their own.  What if the flag represents a
valid possible number to be entered?   Fortunately, C++ has a mechanism
which overcomes the problems associated with flags.  

We can test for the end of input in the following manner:

while ( cin >> number )
 {
   do some more processing;
 }


The value of an extraction operator is true provided the extraction
is successfull (i.e. a number is typed in ).  When the input stream has
been exhausted (tha tis there are no more values to read in) the value of
the extraction operation is false and the while loop will terminate.

The expression cin >> number is true if and only if a number has been
read in.  

To indicate that there are no more values to be entered, the user
types in an escape sequence which is specific to the operating system.
In UNIX, the end-of-file escape sequence is the Control-D sequence
(press the Control kety then the D key).   

The program is as follows:

// Author: Ted Obuchowicz
// Jan. 21, 2002
// example program illustrating use of
// while loop to find sum and average of
// from keyboard using control-d to indicate end of file



#include #include using namespace std;



int main()
{

float sum;
float average;

float number;
int number_of_items_processed;

// perform some initialization

sum = 0.0;
number_of_items_processed = 0;

cout << "Please enter a number " << endl;

while ( cin >> number )
{
  
  sum = sum + number;  // update running sum
  number_of_items_processed++; // increment number of items processed so far
  cout << "Please enter a number " << endl;
  
}



if (number_of_items_processed != 0)
{
  average = sum / number_of_items_processed;
  cout << "The sum is " << sum << endl;
  cout << "The average is "  << average << endl;
}
else
{
  cout << " No numbers were entered " << endl;
  cout << " Sum is " << sum << endl;
  cout << " Impossible to compute average... sorry " << endl;
}



return 0;
}


Here is another example of a while loop. The program uses the simple
"keep subracting one number from another until you can't subtract anymore
" algorithm to perform simple integer division which yields an integer
quotient and remainder:

// Author: Ted Obuchowicz
// Jan. 16, 2002
// example program illustrating use of
// while loop to perform simple integer division which
// yields an integer quotient and an integer remainder
// using the grade school division algorithm



#include #include using namespace std;



int main()
{

int dividend;
int divisor;

int quotient = 0;
int remainder;

// read in the two values 

cin >> dividend >> divisor;

cout << "Dividend = " << dividend <<  " Divisor = " << divisor << endl;

// check for a zero divisor

if ( divisor == 0 )
   cout << "Cannot divide by zero you bonehead!!!!" << endl;
else
  {
    while (dividend >= divisor)
       {
         quotient++;
         dividend = dividend - divisor;
       };
    remainder = dividend;
    cout << "Quotient = " << quotient <<  " Remainder = " << remainder << endl;
   }



return 0;
}



The for loop
-----------

C++ has another looping construct which is frequently used.
We will illustrate its use with  a simple program which
performs integer multiplication by repeated addition. The program is based
on the fact that a multiplication of the form:

product = a * b 

is equivalent to adding a to itself a total of b times 
(or adding b to itself a total of a times).  In the early
times of computing (when there were no built-in multiply instructions,
multiplication was actually carried out in such a manner).

// Author: Ted Obuchowicz
// Jan. 21, 2002
// example program illustrating use of
// for loop in a multiplication by repeated addition program
// we will keep it simple and assume only positive operands


#include #include using namespace std;



int main()
{

int product = 0; // initialize to 0 
int multiplicand;
int multiplier;

cout << "Enter the multiplicand " << endl;
cin >> multiplicand ;

cout << "Enter the multiplier " << endl;
cin >> multiplier;

for (int i = 0 ; i < multiplier ; i++ )
{
    product = product + multiplicand;
}

cout << multiplicand << " * " << multiplier << " = " << product << endl;


return 0;
}

The general form of a for loop is the following:

for ( initialization ; test_expression ; postexpression )
    action;

(again as in the case of an if, if more than one statement
 constitutes the action portion, then they must be enclosed
in the {   }  ) : 


for ( initialization ; test_expression ; postexpression )
{
  action1;
  action2;
  ...
   actionn;
}



When a for loop is encountered in a program the following steps
are performed:

1) the initialization is performed first
2) next, the test_expression is evaluated, if the expression is true then
   the step(s) in the Action portion are executed and then postexpression
   is executed
3) the test_expression is reevaluated, if true, the action statements are
   again executed, if false the for loop TERMINATES and control continues
   with the next statement in the program.


Another example:
---------------

We can re-write our while loop program to use a for loop:

// Author: Ted Obuchowicz
// Jan. 21, 2002
// example program illustrating use of
// for loop to find sum and average of 5 floating point values read in
// from keyboard



#include #include using namespace std;



int main()
{

float sum;
float average;

const int size_of_list= 5;
float number;
int number_of_items_processed;

// perform some initialization

sum = 0.0;
number_of_items_processed = 0;

for (int i = 0; i < size_of_list ; i++)
{
  cout << "Please enter a number " << endl;
  cin >> number;
  sum = sum + number;  // update running sum
}

average = sum / size_of_list;

cout << "The sum is " << sum << endl;
cout << "The average is "  << average << endl;


return 0;
}

In a for loop , any of the 

initialization
test_expression
postexpression

may be omitted.  This may appear strange , but consider the following
program

// Author: Ted Obuchowicz
// Jan. 21, 2002
// example program illustrating use of
//  an INFINITE for loop



#include #include using namespace std;



int main()
{

for( ; ; )
{
 cout << "This is an infinite for loop " << endl;
}

return 0;  // actually the program will never return untill you kill it with 
           // control-c 

}



The following is also another variation of the above theme:

for( ; true  ; )
{
 cout << "This is an infinite for loop " << endl;
}

As another example of the versatility of the for loop let us first 
rewrite our general purpose sum and average program:

// Author: Ted Obuchowicz
// Jan. 21, 2002
// example program illustrating use of
// for loop to find sum and average of
// from keyboard using control-d to indicate end of file



#include #include using namespace std;



int main()
{

float sum;
float average;

float number;

// perform some initialization

sum = 0.0;

cout << "Please enter a number " << endl;

for ( int number_of_items_processed = 0; cin >> number; number_of_items_processed++ )
{
  
  sum = sum + number;  // update running sum
  cout << "Please enter a number " << endl;
  
}



if (number_of_items_processed != 0)
{
  average = sum / number_of_items_processed;
  cout << "The sum is " << sum << endl;
  cout << "The average is "  << average << endl;
}
else
{
  cout << " No numbers were entered " << endl;
  cout << " Sum is " << sum << endl;
  cout << " Impossible to compute average... sorry " << endl;
}



return 0;
}


While at firsrt glance, this program appears both syntactically and
semantically correct, it DOES NOT EVEN COMPILE!!!! :

Here are the compilation error messages as reported by the
g++ compiler:

ted@brownsugar Programs 8:47pm >g++ -o for_loop_gen_purpose_sum for_loop_gen_purpose_sum.C
for_loop_gen_purpose_sum.C: In function `int main()':
for_loop_gen_purpose_sum.C:40: name lookup of `number_of_items_processed' changed for new ANSI `for' scoping
for_loop_gen_purpose_sum.C:30:   using obsolete binding at `number_of_items_processed'


The problem is related to RULES OF SCOPE.  Basically, the problem is that
the loop index variable number_of_items_processed is known only within the
body of the for loop.  Outside the for loop, this index variable cannot
be accessed.

The solution is to remove this variable declaration from the for loop
initialization statement and put it as part of the main program:

// Author: Ted Obuchowicz
// Jan. 21, 2002
// example program illustrating use of
// for loop to find sum and average of
// from keyboard using control-d to indicate end of file



#include #include using namespace std;



int main()
{

float sum;
float average;

float number;

// perform some initialization
sum = 0.0;

cout << "Please enter a number " << endl;

int number_of_items_processed = 0; 


for ( ; cin >> number; number_of_items_processed++ )
{
  
  sum = sum + number;  // update running sum
  cout << "Please enter a number " << endl;
  
}



if (number_of_items_processed != 0)   // this will not compile since scope rules
                                      // limit visibility of number_of_items_processed
                                      // to body of for loop
{
  average = sum / number_of_items_processed;
  cout << "The sum is " << sum << endl;
  cout << "The average is "  << average << endl;
}
else
{
  cout << " No numbers were entered " << endl;
  cout << " Sum is " << sum << endl;
  cout << " Impossible to compute average... sorry " << endl;
}



return 0;
}




             

Note that in the case that the test_expression evaluates to false
right away, the body of the for loop is never executed:

// Author: Ted Obuchowicz
// Jan. 21, 2002
// example program illustrating use of
// of a for loop whose action stateements are never executed



#include #include using namespace std;



int main()
{

for(int index = 0 ;  ( 1 > 1024); index++ )
{
 cout << "This will not be printed" << endl;
 cout << "Keith is better than Madonna " << endl;
}


return 0;
}


In the above program, the initialization of the loop index
is performed and the expression ( 1 > 1024) is evaluated,
since it evaluates to false, the body of the loop is never
entered and control continues with the next statement (which
is the end of the program....)


Nested loops:
------------

Loops may be nested.  For example;


// Author: Ted Obuchowicz
// Jan. 7, 2002
// example program illustrating use of
// nested for loops



#include #include using namespace std;



int main()
{

for(int out = 1 ; out <= 5; out++)
   for (int in = 1 ; in <= 5 ; in++)
   {
    cout << "out = " << out << " in = " << in << "  " ;
    if (in % 5 == 0 )
      cout << endl;
   }

return 0;
}


The program output is:

out = 1 in = 1  out = 1 in = 2  out = 1 in = 3  out = 1 in = 4  out = 1 in = 5  
out = 2 in = 1  out = 2 in = 2  out = 2 in = 3  out = 2 in = 4  out = 2 in = 5  
out = 3 in = 1  out = 3 in = 2  out = 3 in = 3  out = 3 in = 4  out = 3 in = 5  
out = 4 in = 1  out = 4 in = 2  out = 4 in = 3  out = 4 in = 4  out = 4 in = 5  
out = 5 in = 1  out = 5 in = 2  out = 5 in = 3  out = 5 in = 4  out = 5 in = 5  


In the above example, the statements in the body of the innermost for loop are
repeated for every value of out in the outer for loop: 

First the value of out is set to 1 and then the inner loop is executed
with values of in ranging from 1 to 5, then the outer loop index out is incremented
to 2 and again the inner loop repeats with values of in ranging from 1 to 5.
The process repeats until the outer loop terminated when out becomes equal to
6. 


The break statement:
-------------------

A break statement can be used inside a loop construct to
prematurely terminate the loop. Program execution then
resumes with the statement which follows the looping statement.
The break can be used in any of the C++ looping constructs.

Here is an example of a for loop with a break statement which
forces an exit from the loop when the user enters the value
-999:

// Author: Ted Obuchowicz
// Jan. 7, 2002
// example program illustrating use of
// break statement inside a for loop to cause premature exit 
// from the loop


#include #include using namespace std;



int main()
{

int exit_value = -999;  // suppose we wish to stop processing the loop if user
                        // inputs this value
int input_value;


for ( ; cin >> input_value ; )
    if ( input_value != exit_value)
       cout << "You entered " << input_value << endl;
    else
      {
        cout << "exit_value encountered in input stream... breaking from loop!" << endl;
        break;  // exit from the loop
       }


return 0;
}



The use of break statements inside of loops is somewhat recommended against
as it makes program behaviour difficult to understand.  Sometimes it is
possible to rewrite the loop without using a break as in the following
possible to rewrite the loop without using a break as in the following
alternate version of the above program:

// Author: Ted Obuchowicz
// Jan. 23, 2002
// example program illustrating use of
// loop rewritten without a break

#include #include using namespace std;



int main()
{

int exit_value = -999;  // suppose we wish to stop processing the loop if user
                        // inputs this value
int input_value;


for ( ;  ( (cin >> input_value)  && (input_value != exit_value) ) ; )
       cout << "You entered " << input_value << endl;




return 0;
}


A BREAK STATEMENT MAY ONLY APPEAR IN A SWITCH STATEMENT OR IN
ANY OF THE LOOP STATEMENTS. IT MAY NOT APPEAR ANYWHERE ELSE!

The following program does not compile (since it contains
a break in an if statement which is not allowed in C++):

ted@brownsugar Programs 8:01pm >g++ -o illegal_break illegal_break.C 
illegal_break.C: In function `int main()':
illegal_break.C:30: break statement not within loop or switch

// Author: Ted Obuchowicz
// Jan. 23, 2002
// example program illustrating 
// an illegal use of a break statement



#include #include using namespace std;



int main()
{

int i, j;

cin >> i >> j;

if ( i < 5 )
  {
    cout << "i less than 5" << endl;
    cout << "Have a nice day" << endl;
    if ( j < 10 )
      {
         cout << " j less than 10 " << endl;
         cout << " have a good evening " << endl;
         break; // this break is illegal
       }
     cout << " j is not less than 10 " << endl;
   }
 
return 0;
}


The continue statement:
-----------------------

The continue statement is used ONLY IN LOOPS.  When a continue statement
is encountered in the body of a loop, the REMAINING statements in the
body of the loop are SKIPPED OVER, and a new LOOP ITERATION commences.
Note that that while a continue statement causes the program to skip any
remaining body statements , it does not skip the loop test_expression.

In a for loop, the continue makes the program skip to the 
postexpression part and then it tests the condition.

In a while loop, the continue causes the program to go directly
to the test expression part of the while (condition) .

Here is a program which uses a continue statement within a 
for loop which reads 5 characters and counts the number
of 't' characters which have been entered:

// Author: Ted Obuchowicz
// Jan. 23, 2002
// example program illustrating use of
// a continue statement within a for loop



#include #include