OO Design - Separating Instance-specific functions from class-specific functions

Question

Given a class that is semantically supposed to be an object of a certain type, but also has multiple operations for operating on objects of its own type, what is the best way(pattern?) of organizing the class in this type of scenario? I find this to be a common occurrence when a developer creates objects but is thinking with a procedural mindset.

Example:

Class User {
  private m_userData;
  function User() {}
  function GetUserData() {}

  function KillAllUsers(){}
  function MaimAllUsers(){}
}

Answer 1

It's not very clear from your description, but it seems that the "KillAllUsers" and "MainAllUsers" methods are operating on a set of users. I would recommend creating a custom UserCollection that has these methods as instance methods or create them as statics and pass in a collection of users. In modern domain model terms, you'd be dealing with a UserRepository.

Answer 2

I'd rather tend to separate these two purposes into a User and a UserManager class. IMO this provides a clearer separation of what the classes are supposed to do.

Edit: As Dunk correctly observed, the actual name in your application should not be UserManager, but something more descriptive of it's actual purpose. Often times you'll find a design pattern for the functionality you are going to use, so the name of the class will be provided by that pattern. This might lead to names such as UserRepository or UserFactory.

Answer 3

Make an abstraction of the operations you want to perform on your objects, for example by creating an interface. Then you can worry later on how to provide the implementation of the actual logic.

Usually, when dealing with collections of objects of, say, type User, it will not be User itself that will implement it, but another object like UserService for example. If absolutely want it to be available on the class level, you can define a static method that returns the interface type and a default implementation.