Game Development Stack Exchange is a question and answer site for professional and independent game developers. It only takes a minute to sign up.
Sign up to join this community
I don't understand how to use a finite state machine with the entity controlled by the player.
For example I have a Mario style game (2d platform). I can jump, run, walk, take damage, swim, etc. So my first thought was to use this actions as states. But what happens, if you are running when you take damage? Or jumping, taking damage and shooting at the same time?
I just want to add functionalities (actions) to the player in a clean way (not using it for all the actions in the entity update) .
I guess the normal way is to have multiple state machines. This way you can stick to atomic states which often makes your live a lot easier than having to deal with complex states.
For example:
In object oriented designed there is the sate pattern for this. It boils down to having one interface per state machine. And a class per possible state which implements this interface.
To give an example: There would be an interface DamageMode and subclasses NormalDamageMode, InvulnerableDamageMode and InvincibleDamageMode. Whenever the player collidates with an enemy the method DamageMode.collision(enemy) is called. It is handled by the class responsible for the current state.
So if damageMode == NormalDamageMode the player will be damaged and the state variable will be set to InvulnerableDamageMode. In InvulnerableDamageMode nothing happens and in InvincibleDamageMode the enemy will be damaged. Of course there needs to be timers to go back from InvulnerableDamageMode and InvincibleDamageMode to NormalDamageMode.
The main advantage of the state pattern compared to if-elseif-elseif-elseif-blocks is that it allows to structure your code better.
States are not actions, in your game they would control how the character is updated (e.g. each state could have its own update() function).
Take a look at the diagram on this page: http://en.wikipedia.org/wiki/State_diagram. The states tell us something about the world (the door is open or closed). The actions show us what takes the world from one state to another.
In your case you could have states like:
And the transition between states (the events / arrows between states) would be player inputs and reactions to the world, such as:
Each state needs to check if any relevant events have occurred. Typically this is done with if statements.
But what happen when you are running when you take damage?
The hit by monster event would be triggered and transition the player into the in pain state.
or jumping taking damage and shooting at the same time?
The player would transition into the in pain state again which is where you could display your "hit" sprite that shows the character falling backwards.
You probably aren't going to escape if statements, but you can avoid deeply nested statements.
