Refactoring the AI

Today I started this big task… This is the result so far:

Probably something simple to solve, though…

Anyway, I’m replacing the current FSM-based approach to a more organic one, based in an action layer (in the C++ code) and a decision layer (on a Lua script). This way it’s probably easier to handle the multiple enemy states, clearer transitions (a big problem), and a less hacky code for “interruptions of action”.

Oh, and by the way: screenshot number 300! Yay! Celebration!

Link of the Day: This is like QWOP for multiplayer… I’m not seeing myself playing, but at least it’s original: