Which is better OO design

Question

Hi,
I have a class, lets call it MyClass.
public class MyClass {
  private String my_data;
  public MyClass(String s) {
    my_data = s;
  public String toString() {
    return my_data;
}Now, I have a different class that has a member variable called my_class of type MyClass. This other class implements Cloneable. So I write it like this:
  public MySecondClass clone() {
    return new MySecondClass(new MyClass(my_class.toString()));
  }It seems to me there is one other option, and that would be to have MyClass have a constructor that accepts its own type. Then MySecondClass.clone() would look like this:
  public MySecondClass clone() {
    return new MySecondClass(new MyClass(my_class));
  }Is one approach more preferable to the other? I don't want to look like an idiot and pick a bad design method.
Thanks.

Answer 1

The problem with Cloneable, as I understand it, is that there's no ironclad guarantee that the super
class has implemented Cloneable.Yes there is; ad ultimo there's the protected clone() method in the Object class. But there's
some fuzzy voodoo going on in that method, i.e. it checks whether or not the originating class (indirectly)
implemented the Cloneable interface. If not it starts whining, if so, it copies the entire originating
object by shallow copying all its members.
The nasty thing about cloning is that there is no guarantee that an appropriate clone() method
is available even if the class implements the Cloneable interface. It's all up to the responsibility
and discipline of the programmer.
The advantage of cloning is that one doesn't need to know the class an object belongs to, the
disadvantage is that clumsy implementation.
kind regards,
Jos