在我的一次采访中,我被要求解释接口类和抽象类之间的区别。
以下是我的回答:
Methods of a Java interface are implicitly abstract
and cannot have implementations. A Java abstract class can have
instance methods that implements a default behaviour.
Variables declared in a Java interface are by default final. An
abstract class may contain non-final variables.
Members of a Java interface are public by default. A Java abstract
class can have the usual flavours of class members like private,
protected, etc.
A Java interface should be implemented using keyword “implements”; A
Java abstract class should be extended using keyword “extends”.
An interface can extend another Java interface only, an abstract class
can extend another Java class and implement multiple Java interfaces.
A Java class can implement multiple interfaces but it can extend only
one abstract class.
然而,面试官并不满意,他告诉我这种描述代表了“书本知识”。
他让我给出一个更实际的回答,用实际的例子解释我什么时候会选择抽象类而不是接口。
我哪里错了?
接口是一个“契约”,其中实现契约的类承诺实现方法。举个例子,当我将一款游戏从2D升级到3D时,我不得不编写一个界面而不是类。我必须创建一个界面来共享2D和3D版本的游戏类别。
package adventure;
import java.awt.*;
public interface Playable {
public void playSound(String s);
public Image loadPicture(String s);
}
然后我可以实现基于环境的方法,同时仍然能够从一个不知道正在加载的游戏版本的对象调用这些方法。
公共类Adventure扩展了JFrame实现了Playable
公共类Dungeon3D扩展了SimpleApplication实现的Playable
公共类Main扩展了SimpleApplication实现了AnimEventListener
ActionListener,播放
通常,在游戏世界中,世界可以是一个抽象类,在游戏中执行方法:
public abstract class World...
public Playable owner;
public Playable getOwner() {
return owner;
}
public void setOwner(Playable owner) {
this.owner = owner;
}
当我试图在两个密切相关的类之间共享行为时,我创建了一个包含公共行为的抽象类,并作为两个类的父类。
当我试图定义Type(对象的用户可以可靠地调用的方法列表)时,我创建了一个接口。
例如,我绝不会创建一个只有一个具体子类的抽象类,因为抽象类是关于共享行为的。但是我很可能创建一个只有一个实现的接口。我的代码的用户不会知道只有一个实现。实际上,在未来的版本中可能会有几个实现,它们都是一些新的抽象类的子类,这些抽象类在我创建接口时甚至还不存在。
这似乎也有点太书生气了(尽管我在记忆中从未见过这样的说法)。如果面试官(或OP)真的想要更多关于这方面的个人经验,我早就准备好了关于界面是出于必要而进化的轶事,反之亦然。
One more thing. Java 8 now allows you to put default code into an interface, further blurring the line between interfaces and abstract classes. But from what I have seen, that feature is overused even by the makers of the Java core libraries. That feature was added, and rightly so, to make it possible to extend an interface without creating binary incompatibility. But if you are making a brand new Type by defining an interface, then the interface should be JUST an interface. If you want to also provide common code, then by all means make a helper class (abstract or concrete). Don't be cluttering your interface from the start with functionality that you may want to change.