Object-oriented programming

Andrés Cabrera

Functional

Procedural

Object oriented

Paradigm

• Functional: no state
• Procedural: state
• Object oriented: state and functionality handled through "Objects"
• Objects group conceptually and functionally related data and actions

Object

• A data structure (a way to organize data)
• A collection of functions that can be applied to that data

(+ 3 4)

int a = 3;
int b = 4;
int c = a + b;

Numbers n;

n.setNumbers(3,4);

c = n.add();

Functional

Procedural

Object oriented

Class

• A data type
• Defines the "members" of the data structure
• Is the "shape" of an object
• Classes are instantiated as objects

Object

• Each individual instance of a class
• An Object takes up a space in memory in the "shape" of the class
• A class can be instantiated multiple times into multiple objects

Class

• The blueprint

Object

• The house

Class

• Data (called Attributes, fields or properties)
  • Any number of variables
  • Of any type
• Actions/Behavior (functions, called methods)
  • Any number of functions
  • That have access to the attributes
• An abstraction of reality/a bundling of functionality

Why?

• An abstraction of reality/a bundling of functionality
  • This can simplify writing and reading code
• Encapsulation
  • Implementation is separate from interface
  • Irrelevant data and functions can be hidden and protected
• Reuse
  • Using classes can encourage and simplify reuse (debatible...)

Object Oriented languages

• C++
• Java
• Smalltalk
• Python
• C#
• Javascript
• Many others

C++

• C incremented by one
• A super-set of C (mostly)
• Began life as "C with classes" in 1979

A Class in C++

class Number
{
public:
    void setNumbers(int c, int d) {
        a = c;
        b = d;
    }
    int add() {
        return a + b;
    }
private:
    int a;
    int b;
}

A Class in Python

class Number:
    self.a = 0
    self.b = 0

    def setNumbers(self, c,d):
        self.a = c
        self.b = d
    
    def add(self):
        return a + b
}

Static and dynamic types

Number n;
Number m;

n.setNumbers(3.5,4);
m.setNumbers(5,6);

n.add(); // C returns 7, Python return 7.5
m.add(); // Return 11

Objects in C++

Number n;

n.setNumbers(3,4);
int c = n.add();

printf("%i", n.a); // Error a is a private member

Members of an object are accessed using '.'

whether they are methods or attributes

Objects in Python

n = Number();

n.setNumbers(3,4);
c = n.add();

print n.a // Not private!!

Everything is an object in Python!

Constructors and destructors

• Classes often have constructor and or destructor functions
• These are functions that are called automatically when an object is created or destroyed (e.g. a new variable is declared or when it goes out of scope)
• Used to set up initial states and to clean up after use.

class Numbers
{
public:
    Numbers(int a = 0, int b = 0) {
        c = a;
        d = b;
    }
    ~Numbers() {
        // The destructor
    }
    int add() {
        return c + d;
    }
private:
    int c;
    int d;
}

int main(int argc, char *argv[])
{
    Numbers n(3,4);
    int c = n.add();
}

Exercise 1

• Make a class that computes Fibonacci numbers
• The constructor can take the number of numbers to calculate
• It should have member functions ("methods") to set the number of Fibonacci numbers and to print the Fibonacci sequence
• Use the class within the main() function to print the Fibonacci sequence with 5, 10 and 15 numbers.

Inheritance

• A class can inherit from another class
  • In some languages multiple inheritance is allowed: inheriting from two or more classes
• A child class inherits all the members (methods and attributes) of the parent class
• It can then add functionality to the class (more members)
• Or it can override functionality of the parent class (for functions only)

class Shape:
    def __init__(self, length):
        self.type = 'none'
    def area(self):
        return 0
    def getType(self):
        return self.type

class Square(Shape):
    def __init__(self, length):
        self.type = 'square'
        self.length = length

    def area(self):
        return self.length**2

class Circle(Shape):
    def __init__(self, radius):
        self.type = 'circle'
        self.radius = radius
    def area(self):
        return 3.14159 * self.radius**2


c = Circle(2)
print c.area() # prints 12.56636 (4 * pi)


c = Square(2)
print c.area() # prints 4

Public Inheritance

• Essentially, all members keep their "privacy" in the subclass
• Public members stay public
• Private members stay private

Private Inheritance

• Everything from the parent class is essentially inaccessible from the child class

http://en.wikipedia.org/wiki/Inheritance_(object-oriented_programming)

There's a bit more to it:

Why inherit?

• You can include a member of the type inside your class, after all... (called class composition)
• To override functionality
• To construct "is a" relationships (interface inheritance)
• To inherit implementations (to bundle functionality)

Anatomy of a Python Class

• Everything in Python is an object!
• Definition:

  • class ClassName(BaseClass):

  • class ClassName:

• Constructor:

  • def __init__(self, arg1, arg2):

• Members:
  • Attributes:

    • self.AttributeName

  • Methods:

    • def method(self, arg1, arg2 ...):

    • self.method(arg1, arg2)

class Shape:
    def __init__(self, length):
        self.type = 'none'
    def area(self):
        return 0
    def getType(self):
        return self.type

class Square(Shape):
    def __init__(self, length):
        self.type = 'square'
        self.length = length

    def area(self):
        return self.length**2

class Circle(Shape):
    def __init__(self, radius):
        self.type = 'circle'
        self.radius = radius
    def area(self):
        return 3.14159 * self.radius**2


c = Circle(2)
print c.area() # prints 12.56636 (4 * pi)


c = Square(2)
print c.area() # prints 4

Anatomy of a C++ Class

• Definition:

  • class ClassName {}

• Constructor:

  • ClassName(int arg1, ...) {}

• Members and methods:
  • Declared within the function and belonging to private, public or protected sections (private is the default for C++ classes)
  • You can give default parameters in any function
  • You can implement methods (functions) separately by prepending the class name:

    • int ClassName::function(int arg1, ...) {}

class Numbers
{
public:
    Numbers(int a = 0, int b = 0) {
        c = a;
        d = b;
    }
    ~Numbers() {
        // The destructor
    }
    int add() {
        return c + d;
    }
private:
    int c;
    int d;
}

int main(int argc, char *argv[])
{
    Numbers n(3,4);
    int c = n.add();
}

Common Class idioms

• Getter and setter functions:
  • Can be used to validate and transform values before setting private members
• Inheritance or composition?
• One class per file, usually separated by declaration (in a .h or .hpp file) and implementation (in a .cpp file)

Inheritance or Composition

• Some times one might be better than the other
• Often it's just a matter of taste

Exercise 2

• Write a "GoldenRatio" class that has a Fibonacci class variable as a member that offers functions to:
  • Set the number of elements in the Fibonacci sequence to compute
  • return the golden ratio

C++ Namespaces

• Classes and functions can be grouped and organized in "namespaces"
• This is useful to avoid name clashes
• The namespace of a function should be prepended:

  • namespace::function()

• Or the namespace should be brought in to the global namespace:

  • using namespace 
    

C++ templates

• Essentially a way to allow variable types (i.e. types that can vary)
• The type is determined when compiling the program and is not fixed (i.e. each program can choose the type used)
• You set the type for a templated class after its name, within <  and >
• The class will internally use the type you set.
• Writing (well written) templated classes can be tricky
• But you might find yourself using them often

Templates

template <typename Type>
Type max(Type a, Type b) {
    return a > b ? a : b;
}

#include <iostream>

int main()
{
    // This will call max<int> by implicit argument deduction.
    std::cout << max(3, 7) << std::endl;

    // This will call max<double> by implicit argument deduction.
    std::cout << max(3.0, 7.0) << std::endl;

    // This depends on the compiler. Some compilers handle this by defining a template
    // function like double max <double> ( double a, double b);, while in some compilers 
    // we need to explicitly cast it, like std::cout << max<double>(3,7.0);
    std::cout << max(3, 7.0) << std::endl; 
    std::cout << max<double>(3, 7.0) << std::endl;
    return 0;
}

From: https://en.wikipedia.org/wiki/Template_%28C%2B%2B%29

C++ streams

• C++ introducess the streaming operators: >> and <<
• Streaming is similar to setting but depending on context it can mean:
  • append
  • send
  • copy
  • etc.

#include <iostream>

using namespace std;

int main(int argc, char *argv[])
{
    cout << "Hello World" << endl;
    return 0;
}

C++11

• Version of C++ standardized in 2011
• Many significant improvements
• Has been quickly adopted by most compilers
• But sometimes requires additional settings to build on certain platforms

Things I haven't talked about:

(But that are important and you will find mentioned often)

• Operator overloading
• Iterators
• Virtual functions
• Protected functions

Some terminology

API

• Application program interface
• Essentially: the documentation for a library or module
• The public members of a shared class

e.g. https://docs.python.org/2/library/urllib2.html

https://docs.python.org/2/library/json.html

Web API

• API for exchange of information through the web

e.g. Twitter web API

SDK

• Software development kit
• A bundle of software, libraries and documentation
• Targeted to a specific task, device, hardware etc.