Why an abstract class implementing an interface can miss the declaration/implementation of one of the interface's methods?

A curious thing happens in Java when you use an abstract class to implement an interface: some of the interface's methods can be completely missing (i.e. neither an abstract declaration or an actual implementation is present), but the compiler does not complain.

For example, given the interface:

public interface IAnything {
  void m1();
  void m2();
  void m3();
}

the following abstract class gets merrily compiled without a warning or an error:

public abstract class AbstractThing implements IAnything {
  public void m1() {}
  public void m3() {}
}

Can you explain why?

That's because if a class is abstract, then by definition you are required to create subclasses of it to instantiate. The subclasses will be required (by the compiler) to implement any interface methods that the abstract class left out.

Following your example code, try making a subclass of AbstractThing without implementing the m2 method and see what errors the compiler gives you. It will force you to implement this method.

Perfectly fine.
You can't instantiate abstract classes.. but abstract classes can be used to house common implementations for m1() and m3().
So if m2() implementation is different for each implementation but m1 and m3 are not. You could create different concrete IAnything implementations with just the different m2 implementation and derive from AbstractThing -- honoring the DRY principle. Validating if the interface is completely implemented for an abstract class is futile..

Update: Interestingly, I find that C# enforces this as a compile error. You are forced to copy the method signatures and prefix them with 'abstract public' in the abstract base class in this scenario.. (something new everyday:)