With the final hours ticking down before scores are revealed, I thought I’d fill the time by doing a post-mortem for my LD48 compo entry. Revolver is an action puzzle game about growing plants by rotating the planet toward favorable weather conditions. You can play it here!
What Went Right
A:\> Let’s build a time machine!
Time management. Unlike most other jams, the first time ever I felt in control of my time. I didn’t get in everything I wanted to, but I delivered the key elements I needed to. I chose an idea with core mechanic that was design-complete. The idea hit both theme targets. By the end of Saturday, the sandbox prototype was playable, which is a target I don’t always hit. I did a really good job of “layering” my needs, ensuring that every element of the game had a placeholder first, before several iterations of refinement. I don’t just mean art, I also mean text, sound, and code. In this way, as time became short, I could say “you know, that placeholder here is good enough” and polish where it is more needed.
Challenge. I tweaked the levels to be tough on the timer, particularly the last 3-4 levels. While this got some users cursing at me, most said it was on the good side of challenging.
B:\> I’m not looking for judgement, just a yes or no — can you assimilate a giraffe?
Unity. I entered previous compos using Flaxen, which is built on HaxePunk using Haxe. It’s great stuff, but I’ve been wanting to get more experience with a game engine that has 3D, a lighting model, and a proper scene graph. (Hopefully, HaxePunk 3.0 will have these things.) I had a really good time using Unity. I’m not so much of a Microsoftie so was I surprised to enjoy many of C#’s features. I mean I’m never going to love capitalized function names, and Unity/MonoDevelop makes it excruciating to use third party libraries (see how haxelib does it), but overall it was smooth.
In-editor configuration. As much as possible, I exposed all level data and configuration to the editor. Each level was a prefab that I dragged to the Level Manager array in the order I wanted them. This made it easy to tweak and change the levels and messages. Unity does have a really ugly built-in array management with no drag and drop reordering, but it’s passable, and you can customize the inspector view to some extent with a little more time. The configuration for the item types were also fully exposed.
C:\> Good job doing basically nothing
In-game tutorial. Not only did I fit several levels, I also worked in a couple tutorial levels to explain the controls and concepts. I was very pleased to fit this in.
Unfiltered sarcasm. Because I layered my time, I used placeholders everywhere first, and this included the level message texts. The requirement for a placeholder is BUILD IT FAST, so I did what comes to me naturally. That is, I was a sarcastic and obnoxious ass. As time wound down, other things took up my time, and I never went back to the level text. As it happened, however, my sarcasm turned out to be a popular aspect of the game. So yay me and my immaturity!
What Went Wrong
D:\> Please, sir? Can we have some more, please?
Moar levels. I put in just enough levels to introduce the three types of atmospherics (rain, snow and tornado) and three plants (smirkflower, smeggplant, and flurp trees). Barely enough time to dip your toes in.
Less samey. I envisioned a more complex dynamic between the atmospherics and plants. Although I exposed a lot of the configuration to the editor, I didn’t tweak them much. In the end, each plant had an atmosphere it grew 4X as fast in, and an atmosphere that stunted it’s growth completely to 0X. These values could have been tweaked to be more interesting, but with as few levels as I had in there, it would be an underused subtlety.
Moar items. I also pictured two other atmospherics (flock of birds, swarm of bees) that could move around the planet – this would have looked cool! It would have also fed into new asexual and sexual plants which spread via seeds and/or pollen, and were endangered by birds.
Moar animations. My biggest regret on the art side was not making time for a better plant-growth effect. It just scales out along the length, which kind of looks like growth, and kind of looks like it’s rising from a deep bow. What I really wanted was a dynamic art that could support a tweakable growth pattern. If I did that, maybe I could have fit in more plants.
Moar UI. The level transitions were very minimalist. The font was hard to read on some levels if there was foul weather near the south pole.
Overall I had a great time this Ludum Dare. My feedback has been largely positive. Regardless of my scores, I’m still very pleased with the result.