When is it better to use a condition-controlled loop rather then a count-controlled loop?

Answer 1

All loops must be condition-controlled with at least one terminating condition that is reachable. The two main types of loop are called counted loops and infinite loops.

Counted loops are typically implemented using a for statement. Although you can use while, do-while or goto loops to achieve the same thing, we must always strive to express our ideas directly in code and the for loop directly expresses the notion of a counted loop:

for(int i=0; i<100; ++i) {

// ...

}

Here, the body of the loop will execute 100 times. There is no need to comment the loop to state this fact because the for statement contains everything we need to know about this loop; it's all up front at the start of the loop. The initialiser (int i=0) executes one-time only. The conditional expression (i<100) is evaluated before each iteration and the action (++i) is executed at the end of each iteration.

Each of the three arguments in a for statement is optional, thus it is the most versatile of the looping statements. We can also use the comma-operator to perform two or more actions at the end of each loop if we need to. If we leave out all the arguments then we create an infinite loop:

for (;;) {

//...

}

A for statement with no arguments is simply taken to mean "for ever" (or forever). However, the while loop arguably expresses the notion of an infinite loop better:

while (1) {

// ...

}

Of course, even with an infinite loop we still need a terminating condition otherwise our code would literally get stuck in the loop forever. We can place break statements anywhere in the body, however terminating conditions are ideally placed up front at the start of the loop where they are easy to see:

while (1) {

if (!condition1) break; if (!condition2) break;

if (!condition3) break;

// ...

}

In this case it would arguably make far more sense to place the conditions in the loop statement itself:

while (condition1 && condition2 && condition3) {

// ...

}

How you structure your loops is entirely up to you, however the aim is to express the logic of the loop as clearly as possible. Placing the logic up-front simply makes your code easier to read and easier to maintain.

A do-while loop differs from a for loop and a while loop in that the condition is always evaluated at the end of each iteration. Thus the loop body will always execute at least once. If you need to use this looping structure, it helps to comment the start of the loop with the terminating condition in order to help the reader:

int i = random();

do { // until i is 100

// ...

// ... lots of code...

// ...

} while (i!=100);

We can also use goto loops to create highly-complex looping structures with multiple entry and exit points however these are best avoided as they are difficult to read. Anything that is difficult to read is also difficult to maintain, but goto loops can turn even the simplest code into incomprehensible "spaghetti" code. The only practical use of a goto statement is in order to break out of a nested loop:

for (int x=0; x<10; ++x)

{

for (int y=0; y<10; ++y)

{

if(x*y>50) goto exit_door;

// ...

}

}

exit_door:

// ...