New Project Updates and Burn-Out
Intro
I have mostly finished the AI system for the guards in the game however it ended up not being the GOAP system that I initially thought I would need. I ended up switching to a Finite State Machine for a number of reasons of which the main one being that it ended up being a too complex system for the type of behavior I wanted. Another thing I have experienced over the last couple weeks is a severe sense of burn out. To be honest I didn't work on the game much of this week and only felt the motivation to work on it the last day or 2.
New Project Updates
Let's start with the updates. I initially started developing the GOAP system however when I began to implement it
and set up behaviors I found that I wanted to interject into them and more predictably change when the agent would
do certain behaviors. While this is doable in a GOAP system I figured that a more simpler approach would be a Finite
State Machine(FSM). I ended up scrapping the GOAP system in favor of the FSM and I found a lot of the behaviors
much easier to implement how I wanted. Additionally, since this game is within the puzzle genre, I wanted the behaviors
to be a little more predictable so the player could anticipate them more easily. Currently, the Guard will start by
investigating the start position of the Player (as if the Player made a noise when breaking into the building) and if
they don't see anything they will immediately go off and check to see if the treasure is still there. I'm still working
on the behaviors after that but the underlying structure is there. I found a great resource called UnityHFSM which I will
link below with the Pull Request link. It was fairly simple to implement after an hour or so of playing around with it and
I was able to put in a simple few states and transitions.
The main issue that I ran into was getting the Guard to detect the various spots that they were going towards. How the
current Guard is configured is that they have a flashlight that covers the first 2 spots in front of them and the 2 spots
adjacent to them. The Guard is considred to be "seeing" something when that something is colliding with the flashlight. The
current flashlight is formed by creating a mesh out of raycasts outputted from the center point of the Guard in a cone shape.
Initially I tried to detect objects through those raycasts however using the default Raycast() method would only detect
the first hit GameObject and using RaycastAll() wouldn't allow for me to have the flashlight collide properly with the world.
Looking back on it I could have used the array generated by RaycastAll() to detect the closest point hit and stop it there
however, my solution was to just do a RaycastAll() in addition to the normal Raycast() to detect all of the various
GameObjects within the field of view. This ended up working really well and for the time being I think I will stick with this
method.
Guard Set-Up
Burn-Out
Over the last couple weeks I have been experiencing something that I heard a ton about at GDC and have heard about a ton from others in the industry: Burn-Out. While I am definitely no stranger to burn-out, I can definitely feel myself becoming less motivated to apply to jobs and work on projects. As it stands now, I have been searching for a job for the last 2 years. Over the last 6 months it has seemed like all of the jobs I apply for either ghost me or I get rejected. I know that I'm qualified for the positions and I have gotten interviews in the past but it gets to be very draining. I've started branching out into different fields around 6-8 months ago but unfortunately it seems like that's not paying out either. I think the main thing is that I'm just very tired. Right now I'm trying to find a job in retail so I can have more to break up my day and make more money than I do now but we'll see how that goes in the coming months. For now I'm still applying to Games and Software/Web jobs everyday and I want to still work on projects although I'm going to try and take it easy on myself the next week or so.
Next Sprint: AI and Game Loop
The next sprint is mainly going to be focused on finishing up the AI systems and adding in some sort of Game Loop sequence where the Player can get the treasure and get back to the entrance without getting caught by a Guard. I'm pretty confident that I can finish it in two weeks as most of the underlying systems are already there I just need to write some more behaviors and general game logic for a few triggers.