6.

096 Problem Set 3

1 Additional Material
1.1 Arrays of class objects
An array of class objects is similar to an array of some other data type. To create an array
of Points, we write
Point parray [4];
To access the object at position i of the array, we write
parray [ i ]
and to call a method on that object method, we write
parray [ i ]. methodName ( arg1 , arg2 , ...) ;
To initialize an array of objects whose values are known at compile time, we can write
Point parray [4] = { Point (0 ,1) , Point (1 ,2) , Point (3 ,5) , Point (8 ,13) };
We can also allocate an array of objects dynamically using the new operator (this implicitly
calls the default constructor of each new Point):
Point * parray = new Point [4];

1.2 Static members and variables

Static data members of a class are also known as “class variables,” because there is only one
unique value for all the objects of that class. Their content is not different from one object
of this class to another.
For example, it may be used for a variable within a class that can contain a counter with
the number of objects of the class that are currently allocated, as in the following example:
1 # include < iostream >
2
3 using namespace std ;
4
5 class CDummy
6 {

 7 public :

8 static int n ;

9 CDummy () { ++ n ; }

10 ~ CDummy () { --n ; }
11 };
12
13 int CDummy :: n = 0;
14
15 int main ()
16 {
17 CDummy a ;
18 CDummy b [5];
19 CDummy * c = new CDummy ;
20 cout << a . n << " \ n " ; // prints out 7
21 delete c ;
22 cout << CDummy :: n << " \ n " ; // prints out 6
23 return 0;
24 }

In fact, static members have the same properties as global variables, but they can only be
referenced via the class: either in class methods, via a class instance (someObject.staticVariable,
or via the className::variable construct.
Because these variables are global, if we were to initialize them in a header file we could
end up with that initialization being compiled multiple times (once per time we include the
header). To avoid this, we only include a static member’s “prototype” (its declaration) in
the class declaration, but not its definition (its initialization). This is why line 13 above is
necessary, and why if we were to provide a header file for CDummy, we would still need to put
line 13 in a separate .cpp file. If you get linker errors saying a static int is undefined,
check to see whether you’ve included a line like line 13 in a .cpp file.
Classes can also have static member functions – that is, member functions which are
associated with the class but do not operate on a particular class instance. Such member
functions may not access non-static data members. For instance, we might replace CDummy
above with the following class definition:
1 class CDummy

2 {

3 private :

4 static int n ;

5 public :

6 CDummy () { ++ n ; }

7 ~ CDummy () { --n ; }

8 static int getN () { return n ;}

9 };

getN could then be called as c->getN() or CDummy::getN().

2
 1.3 const member functions
It is clear what const-ness means for a simple value like an int, but it is not clear what
functions should be available on a const object, since functions may allow modifications in
subtle ways that ought to be forbidden on const objects. To specify to the compiler that
a given member function is safe to call on const objects, you can declare the function with
the const keyword. This specifies that the function is a “read-only” function that does not
modify the object on which it is called.
To declare a const member function, place the const keyword after the closing paren­
thesis of the argument list. The const keyword is required in both the prototype and the
definition. A const member function cannot modify any data members or call any member
functions that aren’t also declared const. Generally, const member functions should return
const values, since they often return references/pointers to internal data, and we wouldn’t
want to allow someone to get a modifiable reference to the data of a const object.
1 const string & Person :: getName () const {

2 return name ; // Doesn ’t modify anything ; trying to modify a

3 // data member from here would be a syntax error
4 }

If an object of class Person would be declared as const Person jesse;, no non-const

member functions could be called on it. In other words, the set of const member functions
of a class defines the set of operations that can be performed on const objects of that class.

1.4 String objects

Manipulating strings as char * or char[] types tends to be unwieldy. In particular, it is
difficult to perform modifications on strings that change their length; this requires reallo­
cating the entire array. It is also very difficult to deal with strings whose maximum length
is not known ahead of time. Many C++ classes have been created to solve such problems.
The C++ standard library includes one such class, appropriately called string, defined in
header file string under namespace std.
The string class allows us to do all sorts of nifty operations:
1 # include < string >

2 ...

3 string s = " Hello " ;

4 s += " world ! " ;

5 if ( s == " Hello world ! " ) {

6 cout << " Success ! " << endl ;

7 }

8 cout << s . substr (6 , 6) << endl ; // Prints " world !"

9 cout << s . find ( " world " ) ; // ( prints "6")

10 cout << s . find ( ’l ’ , 5) ; // ( prints "9")

A line-by-line description of the string features this code demonstrates:

3
 3. We can set strings to normal char *’s.

4. We can use the + operator to append things to a string. Don’t worry about how this
works for now; we’ll see in Lecture 9 how to allow your classes to do things like this.)

5. We can use the == operator to test whether two strings are the same. (If we tried to
do this with char *’s, we’d just be checking whether they point to the same string
in memory, not whether the pointed-to strings have the same contents. To check for
string equality with char *’s, you need to use the function strcmp.)

8. We can get a new string object that is a substring of the old one.

9. We can find the index a given string within the string object.

10. We can find a character as well, and we can specify a starting location for the search.

Take a few minutes to play around with the string class. Look at the documentation
at http://www.cplusplus.com/reference/string/string/. In particular, be sure to understand
the behavior of the substr function.

1.5 Type Conversions and Constructors

Any time you call a function, the compiler will do its best to match the arguments you
provide with some function definition. As a last-ditch strategy, it will even try constructing
objects for you.
Say you have a function f that takes a Coordinate object, and that the Coordinate
constructor is defined to take one double. If you call f(3.4), the compiler will notice that
there is no f that takes a double; however, it will also see that it can match the f that
it found by converting your argument to a Coordinate object. Thus, it will automatically
turn your statement into f(Coordinate(3.4)).
This applies to constructors, as well. Say you have a Point class, whose constructor
takes two Coordinates. If you write Point p(2.3, 0.5);, the compiler will automatically
turn your statement into Point p(Coordinate(2.3), Coordinate(2.5);.

1.6 Sources
• http://www.cplusplus.com/

 2 Catch that bug

In this section, the following snippets will have bugs. Identify them and indicate how to
correct them. Do these without the use of a computer!

2.1
1 ...
2 class Point
3 {
4 private :
5 int x , y ;
6
7 public :
8 Point ( int u , int v ) : x ( u ) , y ( v ) {}
9 int getX () { return x ; }
10 int getY () { return y ; }
11 void doubleVal ()
12 {
13 x *= 2;
14 y *= 2;
15 }
16 };
17
18 int main ()
19 {
20 const Point myPoint (5 , 3)
21 myPoint . doubleVal () ;
22 cout << myPoint . getX () << " " << myPoint . getY () << " \ n " ;
23 return 0;
24 }

2.2

1 ...
2 class Point
3 {
4 private :
5 int x , y ;
6
7 public :
8 Point ( int u , int v ) : x ( u ) , y ( v ) {}
9 int getX () { return x ; }
10 int getY () { return y ; }

11 void setX ( int newX ) const { x = newX ; }

12 };
13
14 int main ()
15 {
16 Point p (5 , 3) ;
17 p . setX (9001) ;
18 cout << p . getX () << ’ ’ << p . getY () ;
19 return 0;
20 }

2.3

1 ...

2 class Point

3 {

4 private :

5 int x , y ;

7 public :

8 Point ( int u , int v ) : x ( u ) , y ( v ) {}

9 int getX () { return x ; }

10 int getY () { return y ; }

11 };
12
13 int main ()
14 {
15 Point p (5 , 3) ;
16 cout << p . x << " " << p . y << " \ n " ;
17 return 0;
18 }

2.4

1 ...
2 class Point
3 {
4 private :
5 int x , y ;
6
7 public :
8 Point ( int u , int v ) : x ( u ) , y ( v ) {}
9 int getX () { return x ; }

10 void setX ( int newX ) ;

11 };

12

13 void setX ( int newX ) { x = newX ; }

14

15 int main ()

16 {

17 Point p (5 , 3) ;

18 p . setX (0) ;

19 cout << p . getX () << " " << " \ n " ;

20 return 0;

21 }

2.5

1 ...

2 int size ;

3 cin >> size ;

4 int * nums = new int [ size ];

5 for ( int i = 0; i < size ; ++ i )

6 {

7 cin >> nums [ i ];

8 }

9 ... // Calculations with nums omitted

10 delete nums ;
11 ...

2.6

1 class Point
2 {
3 private :
4 int x , y ;
5
6 public :
7 Point ( int u , int v ) : x ( u ) , y ( v ) {}
8 int getX () { return x ; }
9 int getY () { return y ; }
10 };
11
12 int main ()
13 {
14 Point * p = new Point (5 , 3) ;

7
15 cout << p - > getX () << ’ ’ << p - > getY () ;
16 return 0;
17 }

(Hint: this bug is a logic error, not a syntax error.)

3 Point
For the next several problems, you should put your class definitions and function proto­
types in a header file called geometry.h, and your function definitions in a file called
geometry.cpp. If your functions are one-liners, you may choose to include them in the
header file.
In this section you will implement a class representing a point, appropriately named
Point.

3.1 Foundation
Create the class with two private ints. Name them x and y.

3.2 Constructors
Implement a single constructor that, if called with 0 arguments, initializes a point to the
origin – (0, 0) – but if called with two arguments x and y, creates a point located at (x, y).
(Hint: You will need to use default arguments.

3.3 Member Functions

Support the following operations using the given function signatures:

• Get the x coordinate

int Point::getX() const

• Get the y coordinate

int Point::getY() const

• Set the x coordinate

void Point::setX(const int new x)

• Set the y coordinate

void Point::setY(const int new y)

8
4 PointArray
In this section you will implement a class representing an array of Points. It will allow
dynamically resizing the array, and it will track its own length so that if you were to pass it
to a function, you would not need to pass its length separately.

4.1 Foundation
Create the class with two private members, a pointer to the start of an array of Points and
an int that stores the size (length) of the array.

4.2 Constructors
Implement the default constructor (a constructor with no arguments) with the following
signature. It should create an array with size 0.
Implement a constructor that takes a Point array called points and an int called size
as its arguments. It should initialize a PointArray with the specified size, copying the values
from points. You will need to dynamically allocate the PointArray’s internal array to the
specified size.

PointArray::PointArray(const Point points[], const int size)

Finally, implement a constructor that creates a copy of a given PointArray (a copy
constructor ).

PointArray::PointArray(const PointArray& pv)

(Hint: Make sure that the two PointArrays do not end up using the same memory for
their internal arrays. Also make sure that the contents of the original array are copied, as
well.)

4.3 Destructors
Define a destructor that deletes the internal array of the PointArray.

PointArray::~PointArray()

4.4 Dealing with an ever-changing array

Since we will allow modifications to our array, you’ll find that the internal array grows
and shrinks quite often. A simple (though very inefficient) way to deal with this without
repetitively writing similar code is to write a member function PointArray::resize(int
n) that allocates a new array of size n, copies the first min(previous array size, n) existing
elements into it, and deallocates the old array. If doing so has increased the size, it’s fine

for resize to leave the new spaces uninitialized; whatever member function calls it will be
responsible for filling those spaces in. Then every time the array size changes at all (including
clear), you can call this function.
In some cases, after you call this function, you will have to subsequently shift some of
the contents of the array right or left in order to make room for a new value or get rid of an
old one. This is of course inefficient; for the purposes of this exercise, however, we won’t be
worrying about efficiency. If you wanted to do this the “right” way, you’d remember both
how long your array is and how much of it is filled, and only reallocate when you reach your
current limit or when how much is filled dips below some threshhold.
Add the PointArray::resize(int n) function as specified above to your PointArray
class. Give it an appropriate access modifier, keeping in mind that this is meant for use only
by internal functions; the public interface is specified below.

4.5 Member Functions

Implement public functions to perform the following operations:
• Add a Point to the end of the array

void PointArray::push back(const Point &p)

• Insert a Point at some arbitrary position (subscript) of the array, shifting the elements
past position to the right
void PointArray::insert(const int position, const Point &p)

• Remove the Point at some arbitrary position (subscript) of the array, shifting the
remaining elements to the left
void PointArray::remove(const int pos)

• Get the size of the array

const int PointArray::getSize() const

• Remove everything from the array and sets its size to 0

void PointArray::clear()

• Get a pointer to the element at some arbitrary position in the array, where positions
start at 0 as with arrays
Point *PointArray::get(const int position)
const Point *PointArray::get(const int position) const
If get is called with an index larger than the array size, there is no Point you can return
a pointer to, so your function should return a null pointer. Be sure your member functions
all behave correctly in the case where you have a 0-length array (i.e., when your PointArray
contains no points, such as after the default constructor is called).

10
4.5.1
Why do we need const and non-const versions of get? (Think about what would happen if
we only had one or the other, in particular what would happen if we had a const PointArray
object.)

5 Polygon
In this section you will implement a class for a convex polygon called Polygon. A convex
polygon is a simple polygon whose interior is a convex set; that is, if for every pair of points
within the object, every point on the straight line segment that joins them is also within the
object.
Polygon will be an abstract class – that is, it will be a placeholder in the class hierarchy,
but only its subclasses may be instantiated. Polygon will be an immutable type – that is,
once you create the Polygon, you will not be able to change it.
Throughout this problem, remember to use the const modifier where appropriate.

5.1 Foundation
Create the class with two protected members: a PointArray and a static int to keep
track of the number of Polygon instances currently in existence.

5.2 Constructors/Destructors
Implement a constructor that creates a Polygon from two arguments: an array of Points and
the length of that array. Use member initializer syntax to initialize the internal PointArray
object of the Polygon, passing the Polygon constructor arguments to the PointArray con­
structor. You should need just one line of code in the actual constructor body.
Implement a constructor that creates a polygon using the points in an existing PointArray
that is passed as an argument. (For the purposes of this problem, you may assume that the
order of the points in the PointArray traces out a convex polygon.) You should make sure
your constructor avoids the unnecessary work of copying the entire existing PointArray each
time it is called.
Will the default “memberwise” copy constructor work here? Explain what happens to
the PointArray field if we try to copy a Polygon and don’t define our own copy constructor.
Make sure that your constructors and destructors are set up so that they correctly update
the static int that tracks the number of Polygon instances.

5.3 Member Functions

Implement the following public functions according to the descriptions:

11
 • area: Calculates the area of the Polygon as a double. Make this function pure virtual,
so that the subclasses must define it in order to be instantiated. (This makes the class
abstract.)

• getNumPolygons: Returns the number of Polygons currently in existence, and can be

called even without referencing a Polygon instance. (Hint: Use the static int.)

• getNumSides: Returns the number of sides of the Polygon.

• getPoints: Returns an unmodifiable pointer to the PointArray of the Polygon.

5.4 Rectangle
Write a subclass of Polygon called Rectangle that models a rectangle. Your code should

• Allow constructing a Rectangle from two Points (the lower left coordinate and the
upper right coordinate)

• Allow construct a Rectangle from four ints

• Override the Polygon::area’s behavior such that the rectangle’s area is calculated by
multiplying its length by its width, but still return the area as a double.

Both of your constructors should use member initializer syntax to call the base-class
constructor, and should have nothing else in their bodies. C++ unfortunately does not
allow us to define arrays on the fly to pass to base-class constructors. To allow using member
initializer syntax, we can implement a little trick where we have a global array that we update
each time we want to make a new array of Points for constructing a Polygon. You may
include the following code snippet in your geometry.cpp file:
1 Point constructorPoints [4];
2
3 Point * u pd a te Co n st r uc to r Po i nt s ( const Point & p1 , const Point & p2 ,
const Point & p3 , const Point & p4 = Point (0 ,0) ) {
4 cons tructorPoints [0] = p1 ;
5 cons tructorPoints [1] = p2 ;
6 cons tructorPoints [2] = p3 ;
7 cons tructorPoints [3] = p4 ;
8 return con struct orPoints ;
9 }

You can then pass the return value of updateConstructorPoints(...) (you’ll need to
fill in the arguments) as the Point array argument of the Polygon constructor. (Remember,
the name of an array of Ts is just a T pointer.)

12

5.5 Triangle
Write a subclass of Polygon called Triangle that models a triangle. Your code should

• Construct a Triangle from three Points

• Override the area function such that it calculates the area using Heron’s formula:
�
K = s(s − a)(s − b)(s − c)
a+b+c
where a, b, and c are the side lengths of the triangle and s = 2
.

Use the same trick as above for calling the appropriate base-class constructor. You should
not need to include any code in the actual function body.

5.6 Questions
1. In the Point class, what would happen if the constructors were private?

2. Describe what happens to the fields of a Polygon object when the object is destroyed.

3. Why did we need to make the fields of Polygon protected?

For the next question, assume you are writing a function that takes as an argument a Polygon
* called polyPtr.

4. Imagine that we had overridden getNumSides in each of Rectangle and Triangle.

Which version of the function would be called if we wrote polyPtr->getNumSides()?
Why?

5.7 Putting it All Together

Write a small function with signature void printAttributes(Polygon *) that prints the
area of the polygon and prints the (x, y) coordinates of all of its points.
Finally, write a small program (a main function) that does the following:

• Prompts the user for the lower-left and upper-right positions of a Rectangle and
creates a Rectangle object accordingly

• Prompts the user for the point positions of a Triangle and creates a Triangle object
accordingly

• Calls printAttributes on the Rectangle and Triangle, printing an appropriate mes­

sage first.

13

6 Strings
In this section you will write a program that turns a given English word into Pig Latin. Pig
Latin is a language game of alterations played in English. To form the Pig Latin version of
an English word, the onset of the first consonant is transposed to the end of the word an an
ay is affixed. Here are the rules:

1. In words that begin with consonant sounds, the initial consonant (if the word starts

with ‘q’, then treat ‘qu’ as the initial consonant) is moved to the end of the word, and

an “ay” is added. For example:

• beast : east-bay
• dough : ough-day
• happy : appy-hay
• question : estion-quay

2. In words that begin with vowel sounds, the syllable “way” is simply added to the end

of the word.

Write a function pigLatinify that takes a string object as an argument. (You may
assume that this string contains a single lowercase word.) It should return a new string
containing the Pig Latin version of the original. (Yes, it is inefficient to copy a whole string
in the return statement, but we won’t worry about that. Also, your compiler is probably
clever enough to do some optimizations.) You may find it useful to define a constant of type
string or char* called VOWELS.
Remember that string objects allow the use of operators such as += and +.
(Your answers for this problem should go in a separate file from geometry.h and geometry.cpp.)

14

MIT OpenCourseWare
http://ocw.mit.edu

6.096 Introduction to C++

January (IAP) 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.