Design Decisions

Since yesterday was spent investigating potential answers to my temporal coupling blog post, today’s post will be a shorter one. A couple of things came up over the weekend that are worth noting though.

The story that I am currently working on is allowing a player to choose whether he/she plays against another Human player or computer opponent. I have decided that the cleanest way to do this is to build a PlayerFactory. I have implemented it below:

public class PlayerFactory {

  public List<Player> create(int option) {
    if (option == 1) {
      return Arrays.asList(new Human(X), new Human(O));
    } else if (option == 2) {
      return Arrays.asList(new Human(X), new DumbComputer(O));
    return new


A couple of discoveries arose in the construction of this class:

  1. There is a very distinct order to the beginning of the game, one that somewhat defines how classes can/should be initialised. Based on my current implementation, before the Game can be initialised it needs two Players, before two Players can be initialised they each need a Mark assigned to them. In order to prompt the user for the game type and ultimately mark choice the UserInterface needs to be initialised with and input and output stream. Finally, the Game currently needs to be running in order to control the execution of the UserInterface. A bit of a chicken and the egg situation which I will try to resolve soon.
  2. I tried very hard to test behaviour rather than implementation, however, I think I have fallen into the same trap as before. I am now testing that the PlayerFactory produces an list with a suitable Player instance in each spot.
  3. In order to do the aforementioned implementation testing, I did get to do something new, that was to overwrite equals and hashCode for the Player implementations. I had come across the requirement in the Java Koans so used that to help me write something. Unfortunately, for whatever reason, my implementation did not work, but it did have a fantastically useful error message:
Expected :java.util.Arrays$ArrayList<[,]>
Actual   :java.util.Arrays$ArrayList<[,]>

I then discovered that IntelliJ can generate an equals and hashCode for you automatically, so used that and it worked perfectly. I will spend sometime Monday investigating why that happened.