redshift

Hi, I’m Charlie. I’m a 14 year old developer, and I just released my first game, Redshift. Redshift is an arcade-y neon shoot ‘em up. In this article I’ll be covering the tools I used to create it.

Love2D

The game uses the Love2D game engine with scripting in Lua. Love2D is great for prototyping. However, it doesn’t encourage good organization which ended up being an issue halfway through development. Lua doesn’t have true object oriented programming, so I ended up repeating myself a lot. The code structure was a horrendous mess of giant tables and lots of for loops. I ended up refactoring the code and used a Lua class library, which helped out with code structure a lot. I still had to use tables to store instances of the classes (simply because of the design of the engine), but I at least had inheritance and mixins, and the resulting code was much cleaner.

Vim

Vim is - in my opinion - the best text editor out there. It’s fast, powerful and has a million different plugins. I’ve put my .vimrc file on GitHub, but I’ll give a quick rundown of the plugins I use.

• CTRL-P: Allows you to fuzzy search for files in subdirectories.
• delimitMate: Simply adds smart bracket completion.
• Nerdcommenter: Powerful comment handling.

Other than the syntax highlighting, I didn’t use anything Lua specific.

Git

I use git for all my projects, and host my files on GitHub (for public projects) or BitBucket (for private projects). There are a few limits on how many users can access the repository if you use a private BitBucket repository, but since I was the only one working on Redshift, it didn’t really matter.

The Good, the Bad and the Ugly.

Since this was my first longterm game project, I didn’t really know of a good plan of attack to follow. I just developed the game by the seat of my pants.

Things that went right

• I’m really glad I decided to use Lua/Love2D for a few reasons:
  • I know them really well, so I didn’t have to learn how to use an engine/language before I even started development.
  • You can develop very quickly with Lua.
  • Love2D has a fairly large community that answered all my questions on the forums.
• I picked a small enough project that I could actually finish development.
• I managed to release version 1.0 (which has never happened before, except with my Ludum Dare submission)

Things that went not so right

• I didn’t plan. This ended up being much more of a problem than I anticipated. Because:
  • I ended up rewriting almost 50% of my codebase half way through the project because of refactoring, or because of deleting features.
  • If I had planned the final result of the game, I could have finished in about 3 months instead of the 7 months it took.
  • 6 months into the development cycle I still couldn’t figure out how to add progression to the game, so it was more of a lame tech demo than a game.
• I was originally going to make it open world, but had to scrap that idea later in development when I started fighting with the engine to try and make it work.

Things that just kind of happened

• I created all the assets (art, sound, music)
• I spent a lot of time coming up with really cool weapons (I didn’t end up implementing them for version 1.0, but they may reappear in later versions.)
• I spent a lot of time writing the AI. This isn’t really good or bad, because I did enjoy writing it, but it did take a while to get it working right.

Suggestions

If you’re an indie, who is trying to release a game, or you’re just interested in making a game, here are a few suggestions based on my experience with Redshift.

• Plan
  You can see how much this screwed me over because I didn’t. I wasted time and energy rewriting code. If I had planned a final result before haphazardly implementing features, I would have had a much smoother time.
• Write good code
  It will seriously save your butt in the long run. Unless you’re really on a time constraint, write solid code.
• Pick an engine and a language that you know, and are easy to work with
  Many of my previously failed projects were because I was trying to make a game while simultaneously tried to learn a game engine. Write a few very simple projects in the engine before you start long term development.
• Create a game you can finish
  Nothing’s more frustrating than getting super excited that you’re going to create the next Call of Duty and burning out when you make no headway. I would recommend sticking to a 2D top down or puzzle game to start. You’re more likely to finish it.
• Polish the game til it shines
  I can’t necessarily say that Redshift is the most polished game ever, but it doesn’t have any glaring bugs or graphics glitches. You’ll be much happier with the result.

Welp, that’s all. I hope this will help a beginner to start out and will warn of some of the big pitfalls I fell in. As it’s my first game, I’ve still got a lot to learn, but I’m still really happy I’ve done it. If you’re interested in buying Redshift, you can use the widget below ;)