I don’t know that I have ever written a blog post about a game I’ve prototyped before. I mean, of course I prototype every game, but I don’t often talk about games that are in the early phase of development. In this case, I spent a non-trivial amount of time working on a game with the WIP title “Snake Shapes” in December 2021. I don’t think I’m going to spend any more time on it unless I have some big epiphany or something. I don’t know how many games I’ve worked on where I abandon them like this, but there have been a few over the years, I just don’t think I’ve done very much, if anything, to document one of them before.
First, what is Snake Shapes?
The original idea came from an entry in my game design journal on October 15th. Around the end of November, I knew I was going to have most of December to spend on some new game idea, and “Snake Shapes” was just one of many ideas that were competing for my attention. It wasn’t actually what I considered “the best idea”, but it was one that I thought would be “quick and easy” to prototype using a new written-in-swift version of my Generic Game Model (GGM) library.
Anyway, the idea for the game was a turn-based version of Snake, where you try and form shapes (initially just squares) with your tail. This evolved during development a bit to have less emphasis on the actual shapes created. I was going to have a pop-up every time you created a shape, saying how many of that shape you’d created, and giving extra points for shapes that hadn’t been created before. I did not implement that aspect. Every time the snake moves there is something new that gets added to the gameboard. The first 8 are fixed pickups of only two colors, but they are randomized after that, and every 4th thing added is a “wall” square that you cannot move into or through. The game as it is currently is just to get the highest score by eating things and making shapes with your tail until you run out of room to move.
I did spend a bit of time trying to brainstorm different names for the game. I might use the terms “Snake Shapes” and “Square Snake” interchangeably, because I’ve called the project both things, and frequently mix them up. I did also try to entice my 11 year old (who is getting quite good at making art in Procreate) to make artwork for the game, but without any luck.
At some point I had and latched onto an idea that each shape created would boost a counter for “number of power-ups” for a given type of collectible on the gameboard. I thought these would be fixed, one per type, but then I couldn’t decide what all the power-ups should do, so I decided to ask the player which one they’d like after each new square color. I implemented 6 power-ups in total, (which only took 3 days or so, including the popup, which took almost a whole day itself). They are:
- sort your tail by color
- remove half the walls on the board
- convert half the pickups on the board to a random color
- remove half the pickups on the board
- clear out a 5×5 grid around your snake head
- push every wall 1 square away from your current position
Unfortunately, one of the power-ups breaks the game by allowing the player to continue to play indefinitely. Can you guess which one?
Updates to Generic Game Model
I have written a bit a few times about Generic Game Model, but I’m especially proud of a talk I gave about the library to the MN Cocoaheads in 2014. That original version of the library was written in Objective-C, and for at least a year or so, I’ve been working periodically on a new version written in Swift.
Because I haven’t really had a specific project to use it on, I’ve mostly been trying out different things as the mood/whim strikes. For instance, at one point I really wanted it to be a swift package, and you can find a version of it on GitHub that kind of works, but I ran into so many problems using it that way (specifically for Square Snakes development, and while the package is “in flux” so to speak) that I just copied all the files in the project. (I still need to copy the files back out of the project into my package again, as there have been a ton of changes and improvements.)
Another thing I spent extensive time working on was deciding how much to embrace SwiftUI for this iteration of GGM. There’s another repository on Github where I decided to ONLY use SwiftUI, and I abandoned that idea too, after numerous headaches. I think I might know enough now to give that version another go, but I have some concerns (possibly unfounded) about performance, particularly in regards to animation. (SwiftUI does have animatable properties, including offsets and frames, but it was unclear how complex animations would be possible, and animations weren’t the only place I was seeing performance issues, but they may have been my fault entirely.)
What did I really accomplish here?
What am I proud of about working on this game? It’s not really the game idea, frankly. I proved it out, but sadly, I think it’s more fun to play without the power-ups. That said, there are a lot of different directions I could take it, and it could be that it just needs some tweaking in one direction or another to be fun. At this time, I just don’t think I’m going to spend the time necessary to figure that out.
Mostly, the tech was fun to build, and I’ll list some aspects of that here:
- SwiftUI – It was surprisingly quick to build out new screens with SwiftUI. I used it to build essentially everything outside of the game grid itself, including a bunch of tutorial screens, and a popup that asks you to choose which power-up you want to associate with the color of square you just made. I’m not sure it would work in a “real” app, but I was pretty happy with the structure of the State object that controls all of what is displayed in the Application.
- Xcode Cloud Build – This was one of the last things I did on the project, but it’s a fun one. It was only a day or so of work to set up the project so that when I push to the GitHub repository, Xcode Cloud will create a new TestFlight build for me using the new changes. It’s kind of amazing that, as a solo-indie developer, I can get continuous integration working on a project in an afternoon.
- A-star pathfinding – This was actually something that came out of doing Advent of Code (adventofcode.com) in December as opposed to the Square Snake project, but if I hadn’t been working on the project, I am certain I wouldn’t have been as excited about the pathfinding problem. I essentially followed the A-Star tutorial over at Red Blob Games, and had it working in only a few hours. I didn’t really have a use for it in Snake Shapes, but did the work to integrate it with my library anyway, and that work paid off a few days later when I implemented the “push every wall 1 square away” power-up. I didn’t use the A-star itself, but used the PriorityQueue struct that was part of the code written for it.
What could I have done better or differently
This wouldn’t be a good retrospective if I didn’t think about what I could have done differently on this project. To be honest, even though I’m abandoning development, I’m fairly happy with how it turned out. Not happy enough with the game itself that I want to spend any more time on it, but happy with the things I learned and my progress toward making my GGM-swift library a usable tool for native iOS game development. I felt pretty “focused”, and even though what I was building meandered around a bit, I’m pretty happy with how on-task I stayed and the personal growth as a result of the project.
Here’s a few (minor) regrets:
- I would like to have spent more time on the Tetronimo aspect of the game. As previously mentioned, at some point I decided it would be “too easy” to make shapes other than squares in your tail, and abandoned that aspect of the game idea. That meant I could be exceedingly lazy with the code to detect “shapes” in your tail, and allowed me to abandon one of the aspects of the game I was looking forward to developing. (I have several Unity prototypes that do Tetronimo shape detection, and I was kind of looking forward to porting that code to GGM.)
- Also previously mentioned, I abandoned SwiftUI for drawing the game squares themselves. Because the library relies on UIKit for this aspect, it means it’s limited to iOS. If I really embraced SwiftUI, it would support iOS, MacOS, tvOS, and possibly watchOS. This is an area I’d still like to spend more time experimenting.
- I had a TODO in the project to flush out the tutorial views in the app with examples. I didn’t do it because it felt like “work” (in a way that working on the game itself didn’t), but also because I wasn’t sure if I was going to change to using images for the game states themselves at some point, and that might have necessitated re-work. Anyway, the tutorial as it exists now is pretty much the worst kind of tutorial in that it’s just a bunch of text you have to read.
Demo time. Here’s a video of a couple of play-throughs of the last version of the game, so you can get a sense of it.