Why do we use memberless interface ?

Question

What is the point of writing an interface without members ?

INamingContainer is one example in .NET Framework. And it's described in MSDN as :

Identifies a container control that creates a new ID namespace within a Page object's control hierarchy. This is a marker interface only.

Is it used for just this kind of blocks :

if (myControl is INamingContainer)
{
    // do something
}

Or are there other advantages of it ?

EDIT : It was called Marker Interface Pattern (thanks Preet)

Answer 1

This is to ensure that you can treat a collection of objects as the same type and then perform some operations on that type (You know that it must inherit from object).

Think of it like attributes. It's type meta data.

You also might want to perform an operation on a known type (e.g. System.Web.UI.Control) but only perform that operation if the type inherits from a certain interface (e.g. INamingContainer).

foreach(var ctrl in this.Controls)
{
    if (ctrl is INamingContainer)
    {
        // Do something on the control
    }
}

Answer 2

Memberless interfaces are used to provide mixin-like capabilities in C#. So given a class A:

class A : B { ... }

You can give it extra functionality (a-la multiple inheritance) by defining an interface IStuff:

interface IStuff {}

then 'implementing' it in A:

class A : B, IStuff { ... }

and then adding the extra features

class Methods {
  public static void Something(this IStuff stuff) {
    // some functionality here
  }
}

Answer 3

Its a marker interface. It can be used to decorate types so that you can find out it the type is correct at run time with out using reflection. we use it to ensure that generic types are correct in the callee.

Answer 4

To confuse pythonistas who think duck typing is cool.