View
219
Download
3
Category
Preview:
Citation preview
Abstract Data Types (ADT) & UML
C++ Class definition & implementation
constructors, accessors & modifiers
overloading operators
friend functions
HW#1 posted – due: Tuesday, 1/22
Quiz Thursday, 1/24
FAQ 5.14 What is an abstraction and why is it important?
FAQ 5.15 Should abstractions be user-centric or
developer-centric?
An abstraction is a simplified vew of an object
in the user’s own vocabulary. In OO and C++,
an abstraction is the simplest interface to an
object that provides all the features and
services the intended users expect.
User-centric. Focus on the user’s point of view.
Computes what each employee should be paid for a day of work. Reads a file containing start and stop times for each employee. Then calculates and saves the pay amounts to another file.
int runPayCalculator (const char csInputFileName[], const char csOutputFileName[])
Sample input:
510 + 24601
990 - 24601
Sample output
24601 96
int runPayCalculator (const char csInputFileName[], const char csOutputFileName[])
Algorithm:
call computeHours and then computeAndWritePay
Data structure:
The number of hours worked by each employee is stored in an array indexed by the possible employee numbers. This array is of size MAX_EMPLOYEE_NUMBER + 1, where MAX_EMPLOYEE_NUMBER is a global constant
In a well-designed modular program, software components should satisfy the following two properties:
1. Each component performs one well-defined task. (i.e. “cohesion”)
2. Each component is as independent as possible from the others. (i.e. loosely coupled”)
1. Easier to understand; little redundant code.
2. Facilitates software reuse.
3. Easier to implement.
4. Easier to test.
5. Easier to modify.
Independence of modules is typically achieved by “information hiding” (which can be achieved by “encapsulation”).
Procedural Abstraction
Use of a function depends on its purpose (what it does) but not on its implementation (how it does it).
FAQ 5.18 What’s the value of separating interface from implementation?
It’s a key to eliminating the ripple effect when a
change is made.
Class designer/implementer
- designs & implements a class
vs.
Client programmer
- uses a class for an application
vs.
End-user
- uses the application
An Abstract Data Type (ADT) is a specification of a set of data and a set of operations that can be performed on the data.
examples:
String Circle
List Dice
Dictionary Song
Student Telephone Directory
Time Complex number
In C++, a class represents an ADT.
An “instance” of a class is a specific object which is created, and the data members are filled in with values (possibly default values). Objects are created with a specialized member function called a “constructor”.
An instance of a class is destroyed (recycled) with a specialized member function called a “destructor”
Circle
constructor(s): Circle(int,int),
Circle(float,int,int);
float computeArea();
float getRadius();
void setRadius(float);
// etc.
Type name
Public
interface
An ADT is a contract between
The interface designer and ...
The coder of a class that implements the interface
Precondition: any assumption/constraint on the method data before the method begins execution
Postcondition: describes result of executing the method
A C++ program is a collection of functions and classes.
A class represents a set of objects that have common properties.
A class is a template for creating objects.
A class represents a type.
Type determines the set of values an object may have.
Type determines the operations that can be performed on those values.
In C++ there are two kinds of types:
Primitive or build-in types
User Defined or class types
A class consists of members
Data members – also called data fields or attributes
Member functions – also called operators, functions or methods
Data members are also sometimes called instance variables because each object (instance of a class) contains them.
Data members may be either primitive or class types.
private instance variables:
private float radius;
private Point center;
public methods:
constructor(s)
accessor methods (get)
mutator methods (set)
float computeArea()
… etc.
Represent a 2D “point”
Data: (x , y) coordinates, integer values
Methods:
create a point with coordinates (0,0)
create a point with coordinates (x,y),
get the x coordinate of a point,
get the y coordinate of a point,
draw a point
erase a point
move a point
etc.
Unified Modeling Language (UML) is a standard diagram notation for describing a class
Instance
of Person
Field
values
Class
name
Field
signatures:
type and name
Method signatures:
name, argument
types, result type
Point
constructor(s): Point(int,int),
Point(); // default (0,0)
int getX();
int getY();
// etc.
Type name
Public
interface
private instance variables:
private int xCoordinate
private int yCoordinate
public methods:
constructor(s)
accessor methods (get)
mutator methods (set)
… etc.
Class members that are declared in the public section of a class definition are accessible to all functions (inside or outside) the class.
Class members that are declared in the private section of a class definition are accessible only to functions that are members of the class.
Generally we want the operators (member functions) visible to the users of the class.
Thus they are declared public.
Generally we want to keep the implementation details (data members) hidden from the users of the class
Thus they are declared private.
A constructor is a member function that initializes the data members of an object when the object is created.
Note the use of
initialization lists
(more efficient than
assignment
statements)
class Point { public: Point(int i, int j) : x(i), y(j) { } Point() : x(0), y(0) { } private: int x; int y; }
A modifier function provides the ability to modify the value of a private data member
void setX(int newX) {
x = newX;
}
An accessor function provides the ability to read the value of a private data member, without changing it (note use of “const”)
int getX() const { return x; }
class Point {
public:
Point(int i, int j) :
x(i), y(j) { }
Point() :
x(0), y(0) { }
int getX() const
{ return x; }
private:
int x;
int y;
}
Client programmer can write:
Point p1(10,30);
Point p2;
int i = p1.getX();
message receiver
(this)
call the method Point::getX()
class of the receiver
no args
A member function definition (implementation) may be included in the class definition.
The compiler can insert the code for the function body where the function is called.
This is known as an inline function.
Use of inline member function is recommended only for the following:
Functions whose body is very small (one or two lines)
● Constructors
● Accessors
● Modifiers
How would you compare two points, p1 and p2.
Define a method to compare their x and y coordinates.p
p1.lessThan(p2)
Overload the operator <
p1 < p2
bool Point::operator< (const Point& other) const { return (x < other.x) || ((x == other.x) && (y < other.y)); }
class Point {
public:
Point(int i, int j) :
x(i), y(j) { }
Point() :
x(0), y(0) { }
int getX() const
{ return x; }
bool operator<(const Point& other) const;
private:
int x;
int y;
}
bool Point::operator<
(const Point& other) const { return // you fill in...
}
Client programmer can write:
Point p1;
cin>>i; cin>>j;
Point p2(i,j);
if (p2 < p1) then cout << “lol”;
class Point {
public:
// other methods as before
std::ostream& operator<<(
std::ostream& os,
const Point& p);
private: // as before
}
ostream& operator<<(
ostream& os,
const Point& p) {
os << //... you fill in
return os;
}
Point p1(10,30);
cout << p1;
message receiver
(this)
problem: The receiver is type ostream
arg
solution: Make this function a “friend”
class Point {
public:
// other methods as before
friend std::ostream&
operator<<(
std::ostream& os,
const Point& p);
private: // as before
}
ostream& operator<<(
ostream& os,
const Point& p) {
os << //... you fill in
return os;
}
Point p1 =
new Point(10,30);
cout << p1;
solution: Make this function a “friend” Gives permission for this function to have complete access to the data members, even though they are private to the class
A friend is an external function or class that is given the same access to the members of a class as if it were a member.
We declare the ostream insertion operator to be a friend
So it can access the data members and insert their string representation into the output stream.
Because this operator’s left-hand operand is an ostream object, thus it cannot be defined as a member of the Point class.
class Point {
public:
// other methods as before
friend std::ostream&
operator<<(
std::ostream& os,
const Point& p);
private: // as before
}
ostream& operator<<(
ostream& os,
const Point& p) {
os << //... you fill in
return os;
}
/* how do these differ? */
Point p1 = new Point(10,30);
Point p2(10,30);
/* how do these differ? */
cout << p1.getX() << endl;
cout << p1;
/* additional examples were
done in class */
class Point {
public:
// other methods as before
bool operator==(const
Point& other) const;
private: // as before
}
bool Point::operator==
(const Point& other) const
{ return // you fill in...
}
Point p3 =
new Point(10,30);
if (p2 == p3)
cout << “equal
points” << endl;
Now that we have implemented operators == and <,
what about <= ?
Reuse code!
1. Point.h header file (declarations & inline code)
2. Point.cpp implementation file (code)
3. main.cpp test driver
g++ *.cpp
% g++ -c *.cpp
% g++ *.o –o testPt
testPt: main.o Point.o
g++ main.o Point.o –o testPt
main.o: main.cpp Point.h Point.o
g++ -c main.cpp –o main.o
Point.o: Point.h Point.cpp
g++ -c Point.cpp –o Point.o
clean:
rm –f *.o testPoint
Makefile Could type this:
or
% make
% ./testPt
% make clean
Recommended