1. 1 Walter Savitch Chapter 3 Procedural Abstraction and Functions That Return a Value Slides by David B Teague, Western Carolina University, A constituent institution of the University of North Carolina Problem Solving with C++ The Object of Programming

2. 2 3.1 Top Down Design l Step wise refinement, also known as divide and conquer, means dividing the problem into subproblems such that once each has been solved, the big problem is solved. l The subproblems should be smaller and easier to understand and to solve than the big problem.

3. 3 l C++ comes with libraries of predefined functions. l How do we use predefined (or library) functions? l Everything must be declared before it is used. The statement #include brings the declarations of library functions into your program, so you can use the library functions. 3.2 Predefined Functions Libraries

4. #include // include declarations of math // library functions #include // include definitions of // iostream objects using namespace std; // make names available int main() { cout << sqrt(3.0) << endl; // call math library function } // sqrt with argument 3.0, send // the returned value to cout 4

5. 5 3.2 Function call l A function call is an expression consisting of a function name followed by arguments enclosed in parentheses. Multiple arguments are separated by commas. l Syntax: FunctionName(Arg_List) where Arg_List is a comma separated list of arguments. l Examples: n side = sqrt(area); n cout <<“2.5 to the power 3.0 is “ << pow(2.5, 3.0);

6. 6 Display 3.1 A Function call //Computes the size of a dog house that can be purchased //given the user’s budget. #include using namespace std; int main( ) { const double COST_PER_SQ_FT = 10.50; double budget, area, length_side; cout << "Enter the amount budgeted for your dog house $"; cin >> budget; area = budget/COST_PER_SQ_FT; length_side = sqrt(area); // the function call

7. 7 Display 3.1 A Function call (b) cout.setf(ios::fixed); cout.setf(ios::showpoint); cout.precision(2); cout << "For a price of $" << budget << endl << "I can build you a luxurious square dog house\n" << "that is " << length_side << " feet on each side.\n"; return 0; }

8. 8 PITFALL:Problems with Library functions l Some compilers do not comply with the ISO Standard. l If your compiler does not work with #include use #include l Similarly, for headers like cstdlib: use stdlib.h, and for cmath, use math.h l Most compilers at least coexist with the headers without the.h, but some are hostile to these headers.

9. 9 3.2 Predefined Functions Type changing functions Question: 9/2 is an int, but we want the 4.5 floating point result. We want the integer part of some floating point number. How do we manage? Answer: Type casting, or “type changing functions”. C++ provides a function named double that takes a value of some other type and converts it to double. Example: int total, number; double ratio; // input total, number winnings = double(total) / number;

10. 10 PITFALL: Integer division drops the fractional part l REMEMBER: integer division returns just the quotient. The remainder (and any fractional part) are dropped. l If the numerator and denominator are both integer values, the division is done as integer arithmetic, dropping the fractional part. Example: 11 / 2 is 5, NOT 5.5 10 / 3 is 3, NOT 3.333 Even if assigned to a double variable: double ratio = 10/3; // assigned value is 3, NOT 3.3

11. 11 3.3 Programmer Defined Functions function definitions l You can define your own functions. l You can put your functions in the same file as the main function or in a separate file. l If you put your function after the function containing a call to your function, or in a separate file, you must put a prototype (defined in the next slide) for your function some place in the file where it is called, prior to the call.

12. 12 3.3 Programmer Defined Functions function prototypes l A function prototype tells you all the information you need to call the function. A prototype of a function (or its definition) must appear in your code prior to any call to the function. l Syntax: Don’t forget the semicolon n Type_of_returned_value Function_Name(Parameter_list); n Place prototype comment here. n Parameter_list is a comma separated list of parameter definitions: type_1 param_1, type_2 param_2, …. type_N param_N l Example : double total_weight(int number, double weight_of_one); // Returns total weight of number of items that // each weigh weight_of_one

13. 13 3.3 Programmer Defined Functions A function is like a small program To understand functions, keep these points in mind: l A function definition is like a small program and calling the function is the same thing as running this small “program.” l A function has formal parameters, rather than cin, for input. The arguments for the function are input and they are plugged in for the formal parameters. l A function of the kind discussed in this chapter does not send any output to the screen, but does send a kind of “output” back to the program. The function returns a return-statement instead of cout-statement for “output.”

14. 14 3.3 Programmer Defined Functions Display 3.3 A function Definition (Slide 1 of 2) #include using namespace std; double total_cost(int number_par, double price_par); //Computes the total cost, including 5% sales tax, //on number_par items at a cost of price_par each. int main( ) { double price, bill; int number; cout << "Enter the number of items purchased: "; cin >> number; cout << "Enter the price per item $"; cin >> price; bill = total_cost(number, price); The function call cout.setf(ios::fixed); cout.setf(ios::showpoint); cout.precision(2); cout << number << " items at " << "$" << price << " each.\n" << "Final bill, including tax, is $" << bill << endl; return 0; }

15. 15 3.3 Programmer Defined Functions Display 3.3 A function Definition (Slide 2 of 2) double total_cost(int number_par, double price_par) The function { heading heading const double TAX_RATE = 0.05; //5% sales tax The function double subtotal; The function definition body subtotal = price_par * number_par; return (subtotal + subtotal*TAX_RATE); }

16. 16 3.3 Programmer Defined Functions Call-by-value Parameters Consider the function call: bill = total_cost(number, price); l The values of the arguments number and price are “plugged” in for the formal parameters. This process is (A precise definition of what “plugged” means will be presented later. For now, we will use this simile.) l A function of the kind discussed in this chapter does not send any output to the screen, but does send a kind of “output” back to the program. The function returns a return-statement instead of cout-statement for “output.”

17. 17 3.3 Programmer Defined Functions Alternate form for Function Prototypes l The parameter names are not required: double total_cost(int number, double price); l It is permissible to write: double total_cost(int, double ); l Nevertheless, code should be readable to programmers as well as understandable by the compiler, so check for readability and chose to use parameter names when it increases readability. l Function HEADERS (in the definition) must use parameter names. (There is an exception to this rule that we will not deal with in this course.)

18. 18 3.3 Programmer Defined Functions Anatomy of a Function Call (part 1 of 3) Review Display 3.3 0 Before the function is called, the value of the variables number and price are set to 2 and 10.10, by cin. 1 The following statement, which includes a function call, begins executing: bill = total_cost(number, price); 2 The value of number (which is 2) is plugged in for number_par and the value of price (which is 10.10) is plugged for price_par: double total_cost(int number_par, double price_par) { const double TAX_RATE = 0.05; //5% sales tax double subtotal; subtotal = price_par * number_par; return (subtotal + subtotal*TAX_RATE); }

19. 19 3.3 Programmer Defined Functions Anatomy of a Function Call (part 2 of 3) The execution produces the following effect: double total_cost(int 2, double 10.10) { const double TAX_RATE = 0.05; //5% sales tax double subtotal; subtotal = 10.10 * 2; return (subtotal + subtotal*TAX_RATE); }

20. 20 3.3 Programmer Defined Functions Anatomy of a Function Call (part 3 of 3) 3 The body of the function is executed, i.e., the following is executed: { const double TAX_RATE = 0.05; //5% sales tax double subtotal; subtotal = 10.10 * 2; return (subtotal + subtotal*TAX_RATE); } 4 When the return statement is executed, the value of the expression after the return is the value returned by the function. In this case, when return (subtotal + subtotal*TAX_RATE); is executed, the value of (subtotal + subtotal*TAX_RATE), which is 21.21 is returned by the function call bill = total_cost(number, price); the value of bill is set to 21.21 when the following statement ends: bill = total_cost(number, price);

21. 21 PITFALL Arguments in the wrong order l When a function is called, C++ substitutes the first argument given in the call for the first parameter in the definition, the second argument for the second parameter, and so on. l There is no check for reasonableness. The only things checked are: i) that there is agreement of argument type with parameter type and ii) that the number of arguments agrees with the number of parameters. l If you do not put correct arguments in call in the correct order, C++ will happily assign the “wrong” arguments to the “right” parameters.

22. 22 3.3 Programmer Defined Functions Summary of Syntax for a Function that Returns a Value. Function Prototype: Type_Returned Function_Name(Parameter_List); Prototype Comment function header Function Definitions Type_Returned Function_Name(Parameter_List) { Declaration_1 Declaration_2... Must include one or Declaration_Last; more return statements. body Executable _1; Executable_2;... Executable_Last }

23. 23 3.4 Procedural Abstraction Principle of Information Hiding l David Parnas, in 1972 stated the principle of information hiding. n A function’s author (programmer) should know everything about how the function does its job, but nothing but specifications about how the function will be used. n The client programmer -- the programmer who will call the function in her code -- should know only the function specifications, but nothing about how the function is implemented.

24. 24 3.4 Procedural Abstraction The Black Box Analogy l Based on the Principle of Information Hiding, the text describes the BLACK BOX analogy. l When using a function, we behave as if we know nothing about how the function does its job, that we know only the specifications for the function. l When writing a function, we behave as if we know nothing but the specifications about how the function is to be used. l In this way, we avoid writing either application code that depends on the internals of the function or writing function code that depends in some way on the internal structure of the application.

25. 25 3.4 Procedural Abstraction The Principle When applied to a function definition, the principle of procedural abstraction means that your function should be written so that it can be used like a black box. This means the user of a function should not need to look at the internals of the function to see how the function works. The function prototype and accompanying commentary should provide enough information for the programmer to use the function. To ensure that your function definitions have this property, you should adhere to these rules: n The prototype comment should tell the programmer any and all conditions that are required of the arguments to the function and should describe the value returned by the function. n All variables used in the function body should be declared in the function body. (The formal parameters are already declared in the function header.)

26. 26 3.5 Local Variables l Variables declared within the body of a function definition are said to be local to that function, or have that function’s block as their scope. l Variables declared within the body of the main function are said to be local to the main function, or to have that function’s block as their scope. l When we say a function is a local variable without further mention of a function, we mean that variable is local to some function definition. l If a variable is local to a function, you can have another variable with the same name that is local to either main or some other function.

27. 27 3.5 Local Variables Global Constants and Global Variables l A named constant declared outside the body of any function definition is said to be a global named constant. A global named constant can be used in any function definition that follows the constant declaration. There is an exception to this we will point out later. l Variables declared outside the body any function is said to be global variables or to have global scope. Such a variable can be used in any function definition that follows the constant declaration. There is also an exception to this we will point out later. l Use of global variables ties functions that use the global variables in a way that makes understanding the functions individually nearly impossible. There is seldom any reason to use global variables.

28. 28 Display 3.11 Global Named Constants (Part 1 of 2) //Computes the area of a circle and the volume of a sphere. //Uses the same radius for both calculations. #include using namespace std; const double PI = 3.14159; double area(double radius); //Returns the area of a circle with the specified radius. double volume(double radius); //Returns the volume of a sphere with the specified radius. int main( ) { double radius_of_both, area_of_circle, volume_of_sphere; cout << "Enter a radius to use for both a circle\n" << "and a sphere (in inches): "; cin >> radius_of_both;

29. 29 Display 3.11 Global Named Constants (Part 2 of 2) area_of_circle = area(radius_of_both); volume_of_sphere = volume(radius_of_both); cout << "Radius = " << radius_of_both << " inches\n" << "Area of circle = " << area_of_circle << " square inches\n" << "Volume of sphere = " << volume_of_sphere << " cubic inches\n"; return 0; } double area(double radius) { return (PI * pow(radius, 2)); } double vo lume(double radius) { return ((4.0/3.0) * PI * pow(radius, 3)); }

30. 30 3.5 Local Variables Call-by-value Formal Parameters are Local Variables l Formal parameters are more than just blanks to be filled in with values from arguments. l We promised a more precise discussion of what “value plugged in for the argument” means. Here is that discussion. l Formal parameters are actually local variables that are initialized by the call mechanism to the value of the argument. They may be used in any way that a local variable can be used. l Display 3.12 (next slide) is an example of a formal parameter used as a local variable.

31. 31 Display 3.12 Formal Parameter Used as Local Variable (Part 1 of 2) //Law office billing program. #include using namespace std; const double RATE = 150.00; //Dollars per quarter hour. double fee(int hours_worked, int minutes_worked); //Returns the charges for hours_worked hours and //minutes_worked minutes of legal services. int main( ) { int hours, minutes; double bill; cout << "Welcome to the offices of\n" << "Dewey, Cheatham, and Howe.\n" << "The law office with a heart.\n" << "Enter the hours and minutes" << " of your consultation:\n"; cin >> hours >> minutes;

32. 32 Display 3.12 Formal Parameter Used as Local Variable (Part 2 of 2) bill = fee(hours, minutes); The value of minutes is not changed by a call to fee. cout.setf(ios::fixed); cout.setf(ios::showpoint); cout.precision(2); cout << "For " << hours << " hours and " << minutes << " minutes, your bill is $" << bill << endl; return 0; } minutes_worked is a local variable initialized to the value of minutes double fee(int hours_worked, int minutes_worked) { int quarter_hours; minutes_worked = hours_worked*60 + minutes_worked; quarter_hours = minutes_worked/15; return (quarter_hours*RATE); }

33. 33 3.5 Local Variables Name spaces Revisited (1 of 2) All our use of namespaces has amounted to #include using namespace std; While this is correct, we are sidestepping the reason namespaces were introduced into C++, though we have done this for good teaching reasons. In short, we have been “polluting the global namespace.” So long as our programs are small, so this is not a problem. This won’t always be the case, so you should learn to put the using directive in the proper place.

34. 34 3.5 Local Variables Name spaces Revisited (2 of 2) Placing a using directive anywhere is analogous to putting all the definitions from the namespace there. This is why we have been polluting the global namespace. We have been putting all the namespace std names from our header file in the global namespace. The rule, then is: Place the namespace directive, using namespace std; inside the block where the names will be used. The next slide is Display 3.13 with the namespace directive place correctly.

35. 35 3.5 Local Variables Display 3.13 Using Name spaces (1 of 2) //Computes the area of a circle and the volume of a sphere. //Uses the same radius for both calculations. #include #include //Some compilers may use math.h instead of cmath. const double PI = 3.14159; double area(double radius); //Returns the area of a circle with the specified radius. double volume(double radius); //Returns the volume of a sphere with the specified radius. int main( ) { using namespace std; double radius_of_both, area_of_circle, volume_of_sphere; cout << "Enter a radius to use for both a circle\n" << "and a sphere (in inches): "; cin >> radius_of_both;

36. 36 3.5 Local Variables Display 3.13 Using Name spaces (1 of 2) area_of_circle = area(radius_of_both); volume_of_sphere = volume(radius_of_both); cout << "Radius = " << radius_of_both << " inches\n" << "Area of circle = " << area_of_circle << " square inches\n" << "Volume of sphere = " << volume_of_sphere << " cubic inches\n"; return 0; } double area(double radius) The behavior of this program is exactly { the same as that of Display 3.11 using namespace std; return (PI * pow(radius, 2)); } double volume(double radius) { using namespace std; return ((4.0/3.0) * PI * pow(radius, 3)); }

37. 37 3.6 Overloading Function Names l C++ distinguishes two functions by examining the function name and the argument list for number and type of arguments. l The function that is chosen is the function with the same number of parameters as the number of arguments and and that matches the types of the parameter list sufficiently well. l This means you do not have to generate names for functions that have very much the same task, but have different types.

38. 38 3.6 Overloading Function Names Display 3.15 Overloading a Function Name (1 of 2) //Illustrates overloading the function name ave. #include double ave(double n1, double n2); //Returns the average of the two numbers n1 and n2. double ave(double n1, double n2, double n3); //Returns the average of the three numbers n1, n2, and n3. int main( ) { using namespace std; cout << "The average of 2.0, 2.5, and 3.0 is " << ave(2.0, 2.5, 3.0) << endl; cout << "The average of 4.5 and 5.5 is " << ave(4.5, 5.5) << endl; return 0; }

39. 39 3.6 Overloading Function Names Display 3.15 Overloading a Function Name (2 of 2) double ave(double n1, double n2) Both these functions have the { same name, but have parameter return ((n1 + n2)/2.0); lists that are have different } numbers of parameters. double ave(double n1, double n2, double n3) { return ((n1 + n2 + n3)/3.0); }

40. 40 3.6 Overloading Function Names Automatic Type Conversion l We pointed out that when overloading function names, the C++ compiler compares the number and sequence of types of the arguments to the number and sequence of types for candidate functions. l In choosing which of several candidates for use when overloading function names, the compiler will choose an exact match if one if available. l An integral type will be promoted to a larger integral type if necessary to find a match. An integral type will be promoted to a floating point type if necessary to get a match.

41. 41 3.6 Overloading Function Names Display 3.16 Overloading a function name (1 of 4) //Determines whether a round pizza or a rectangular pizza is the best buy. #include double unitprice(int diameter, double price); //Returns the price per square inch of a round pizza. //The formal parameter named diameter is the diameter of the pizza //in inches. The formal parameter named price is the price of the pizza. double unitprice(int length, int width, double price); //Returns the price per square inch of a rectangular pizza //with dimensions length by width inches. //The formal parameter price is the price of the pizza. int main( ) { using namespace std; int diameter, length, width; double price_round, unit_price_round, price_rectangular, unitprice_rectangular;

42. 42 3.6 Overloading Function Names Display 3.16 Overloading a function name (2 of 4) cout << "Welcome to the Pizza Consumers Union.\n"; cout << "Enter the diameter in inches" << " of a round pizza: "; cin >> diameter; cout << "Enter the price of a round pizza: $"; cin >> price_round; cout << "Enter length and width in inches\n" << "of a rectangular pizza: "; cin >> length >> width; cout << "Enter the price of a rectangular pizza: $"; cin >> price_rectangular; unitprice_rectangular = unitprice(length, width, price_rectangular); unit_price_round = unitprice(diameter, price_round); cout.setf(ios::fixed); cout.setf(ios::showpoint); cout.precision(2);

43. 43 3.6 Overloading Function Names Display 3.16 Overloading a function name (3 of 4) cout << endl << "Round pizza: Diameter = " << diameter << " inches\n" << "Price = $" << price_round << " Per square inch = $" << unit_price_round << endl << "Rectangular pizza: length = " << length << " inches\n" << "Rectangular pizza: Width = " << width << " inches\n" << "Price = $" << price_rectangular << " Per square inch = $" << unitprice_rectangular << endl; if (unit_price_round < unitprice_rectangular) cout << "The round one is the better buy.\n"; else cout << "The rectangular one is the better buy.\n"; cout << "Buon Appetito!\n"; return 0; }

44. 44 Chapter Summary( 1 of 2) l A good plan of attack on a problem is to decompose the problem into smaller, more accessible problems, the decompose these into still more manageable problems. This is known as Top-Down Design. l A function that returns value is like a small program. Arguments provide input and the return value provides the output. l When a subtask for a program takes some values as input and produces a single value as its only output, then that subtask can be implemented as a function. l A function should be defied so that it can be used as a black box. The program author should know no more than the specification of the client use, and the client author should know nothing more than the specification of the implementation of the function. This rule is the principle of procedural abstraction. l A variable that is defined in a function is said to be local to the function. l Global named constants are declared using the const modifier. Declarations of global named constants are normally placed at the start of a program after the include directives, before the function prototypes.

45. 45 Chapter Summary (2 of 2) l Call by value formal parameters (the only kind we have discussed so far) are local variables to the function. Occasionally a formal parameter may be useful as a local variable. l When two or more function definitions have the same name, this is called function name overloading. When you overload a name, the function definitions must have different numbers of formal parameters or formal parameters of different types.