By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
446,156 Members | 962 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 446,156 IT Pros & Developers. It's quick & easy.

Where to put attribute when using Inheritance

P: n/a
Hello!!

Assume we have one base class called Vehicle and two derived classes called
Car and Bus.
I would be able to call method getName on an object of class Car or Bus and
return back the name that is set for this class. Asking getName on an object
of class Vehicle is of no interest. Assume we have an attribute called name
of type string.

Now to my question do you think it's right to put the name attribute in each
of class Car and Bus. I think so.
I think putting the name attribute in class Vehicle is wrong because the
name is different in class Car and Bus.

The name attribute is set in the c-tor for class Car and Bus.
This getName is made pure virtual so polymorfism can be used to call the
right getName depending of the object type

I'm I right in my thoughts.

Many thanks
//Tony


Jul 23 '05 #1
Share this Question
Share on Google+
3 Replies


P: n/a
Tony Johansson wrote:
Assume we have one base class called Vehicle and two derived classes called
Car and Bus.
Why are you starting another thread instead of continuing in your previous
"Weapon / Winchester / .." one? You're asking same questions. Did you
read the replies to your first post?
[...]

Jul 23 '05 #2

P: n/a
In my opinion you want three things:
- Have Vehicle objects with a private member 'name'
- Have derived objects from Vehicle - which have a name
- No objects of Vehicle can be created

I would solve this by making the constructor of vehicle not accesible
to the outside world and have the name attribute at the highest
possible level, see a quick example below. If you try to add a line
like 'Vehicle v;' to the main() you get a compilation error.

-#include <iostream>
-#include <string>
-
-using namespace std;
-
-class Vehicle
-{
- string m_name;
-protected:
- Vehicle() {};
-public:
- virtual ~Vehicle() {};
- const string& name() {return m_name;}
- void name(const string& n) { m_name = n; }
-};
-
-class Car:public Vehicle
-{
-public:
- Car(const string& n):Vehicle(){ name(n); }
- ~Car() {}
-};
-
-int main()
-{
- Car c("Fiat");
- cout << c.name() << endl;
-
- return 0;
-}

Jul 23 '05 #3

P: n/a

"Tony Johansson" <jo*****************@telia.com> wrote in message
news:lJ*******************@newsb.telia.net...
Hello!!

Assume we have one base class called Vehicle and two derived classes called Car and Bus.
I would be able to call method getName on an object of class Car or Bus and return back the name that is set for this class. Asking getName on an object of class Vehicle is of no interest. Assume we have an attribute called name of type string.

Now to my question do you think it's right to put the name attribute in each of class Car and Bus. I think so.
I think putting the name attribute in class Vehicle is wrong because the
name is different in class Car and Bus.
Nothing prevents you to support a common attribute and a specialized
attribute.

The name attribute is set in the c-tor for class Car and Bus.
This getName is made pure virtual so polymorfism can be used to call the
right getName depending of the object type

I'm I right in my thoughts.

Many thanks
//Tony


Yes, your choice to place the name attribute in each derived class is an
option. Placing a common attribute in the abstract class is another option.
Same goes for an Engine class. Its a different Engine but its still an
Engine.

Since name isn't a clear attribute, lets consider a common license number
attribute for all vehicles. Let Cars have a car-type attribute and Trucks a
truck-classification attribute. I can implement the pure-virtual display()
member function, overload display() in both Car and truck and still call the
abstract display() through an access specifier.

// CarTest.cpp : Defines the entry point for the console application.
//
#include <iostream>
#include <string>
#include <vector>
using std::cout;
using std::endl;

class Vehicle
{
std::string m_license;
protected:
Vehicle(std::string s) : m_license(s) { }
public:
virtual ~Vehicle() { }
virtual void display() const = 0
{
cout << "vehicle license: " << m_license;
cout << "\t";
}
};

class Car : public Vehicle
{
std::string m_type;
public:
Car(std::string s, std::string t) : Vehicle(s), m_type(t) { }
~Car() { }
void display() const
{
Vehicle::display();
cout << "car type: " << m_type;
cout << endl;
}
};

class Truck : public Vehicle
{
std::string m_classification;
public:
Truck(std::string s, std::string t) : Vehicle(s), m_classification(t)
{ }
~Truck() { }
void display() const
{
Vehicle::display();
cout << "truck classification: " << m_classification;
cout << endl;
}
};

int main()
{
std::vector<Vehicle*> v_vehicles;

v_vehicles.push_back(new Car("M1111", "sportscar"));
v_vehicles.push_back(new Truck("G2222", "2 Tons"));
v_vehicles.push_back(new Car("L3333", "coupe"));
v_vehicles.push_back(new Truck("D4444", "4 Tons"));

typedef std::vector<Vehicle*>::iterator ITER;
ITER it = v_vehicles.begin();
for ( it = v_vehicles.begin(); it != v_vehicles.end(); ++it)
{
(*it)->display();
}

for ( it = v_vehicles.begin(); it != v_vehicles.end(); ++it)
{
delete *it;
}

return 0;
}

Jul 23 '05 #4

This discussion thread is closed

Replies have been disabled for this discussion.