Finally Finish Something 2021 Postmortem

Star Fort Defense

FFS 2021 Postmortem

Now that the FFS 2021 game jam is finally over, it's time to look back and reflect on how things went over the course of the past month.

Going into the game jam, I had made several attempts to get this project off the ground, but never had the motivation to finish something even vaguely presentable. I'd spent a fair amount of time working out what I wanted to make, but I hadn't actually made anything. My goal for the FFS 2021 Game Jam was to create a presentable prototype of the game I wanted to create, in a way that allows me to expand it going forward with minimal rewriting. 

The GUI behaved itself well, and hasn't given me any serious problems yet. The most time-consuming part of creating the GUI so far was designing it. GUI elements related to unimplemented features aren't visible or are greyed out, but the framework to display them is fully in place. Starting the project by designing the GUI was probably the best decision I could have made, because it provided guidance for what features needed to be implemented to have something resembling a playable game.

The player-controlled units have also been mostly unproblematic. For a while there was an issue with the attack script, where they would choose a target but do nothing to attack it, but that was resolved by reducing the number of bullet object copies in the game map at any given time.

The biggest issue that arose during development over the past month is the pathfinding algorithm. For whatever reason, the movement-grid functions built in to GameMaker Studio 1.4 do not play nice with movement on three axes or with the approaches I tried to use for selecting a target to move to. Up until I allowed the player-controlled units to shoot at the monsters, they were mostly okay when finding a target, though things didn't quite work as well as I wanted them to, and while seeking their third target, the system would break down and crash almost every time. Once I allowed the player's units to attack, though, the enemy AI panics, crashing to desktop as soon as the first monster detects a player. In response to that, given the time constraints, I disabled the AI script, leaving the monsters to serve as glorified target practice for now.

Next Steps

The next stage of development has to be fixing the pathfinding issues and creating a new AI for the enemy units. Until that's done, it can't really be called a game, because there is no possibility of failure. I'm also going to rework the equipment system and have units display all of their stats when selected, instead of just their health, XP, and level. How long all of this takes depends on how swamped I get with studying and how exasperating the pathfinding problems continue to be.