Inspiration
I’ve always had an interest in text-based games. It’s always felt like these games needed to be much more engaging than their graphically-gifted counterparts because there were fewer pretty things to look at. They had to engage your imagination, or at least your interest, to such an extent that you would continue to play. It originally started when I was learning that I was good at writing (for a small town high school student), and only grew the more I started using, and learning about, computers. Initially, it was just those text-based adventures that described the scene in the room you were in and presented you with a list of options. Then, I discovered roguelikes. These were the some of the most challenging and captivating games I had played and despite how frustrated they would make me, I kept coming back. Life forced me to put things on the back-burner for quite a while but now, secure enough in my career to start prioritizing my own health and family over work, I am drawn back to this genre with an experienced developer’s cant. These are new problems that I’ve not had to solve and while they may not be terribly complex, they are new and intriguing in that they are a definitive shift away from what I have to solve in my day-to-day.
Plotting the course
In an evening of perusing the web for information about procedurally generated mazes and the various algorithms to find the route through said mazes, I stumbled upon RogueBasin and while reading through their articles on some of the major roguelikes, I discovered a series of articles: Roguelike Tutorials. I will be loosely following this series and doing my best to convert the Python to Go. I’ve found https://github.com/Joakker/tcod-go, “a wrapper of the libtcod library,” which I hope will be enough. Who knows, maybe I’ll go down the rabbit hole of porting some of the tools from libtcod into Go as part of this series, or a follow-up series.
A quick note before we dive in: the code for this project in its current state can be found here: https://github.com/cstaaben/erebor
I decided to call my game Erebor, based on the dwarven mountain/fortress/city from J.R.R. Tolkien’s “The Hobbit.” Moria really seemed fascinating, and Angband was already taken. Hopefully no fans of his work are going to metaphorically crucify me for using The Lonely Mountain as my setting!
My first lesson
Now, obviously, I wrote the above sections prior to actually starting the work. But I didn’t feel they constituted a worthy introduction post and, at the time, I didn’t have much to add to flesh it out. My thought was to take notes and write this post as I developed the game. That approach lasted all the way until the end of the first section for the roguelike tutorial.
At first, I had some normal struggles: installing libtcod and SDL2 on a MacBook Pro with an M1 Pro chip was a series of failed experiments until I built them from source, getting pkg-config to look in the right place to find those packages, translating the Python code and libtcod API calls into the Go library equivalents, and following a basic tutorial while organizing the code to allow me to build upon it without crazy amounts of refactoring in the future. But when I reached the first real functionality of the game (moving the character around the screen), the weaknesses it the library become very apparent. I’m not sure what version of libtcod it was last compiled with and what’s changed since but given the lack of updates to that library, I can only assume they did not stay abreast of those changes. Unfortunately, updating what appears to be a mostly abandoned library does not fall under the umbrella of this series (nor do I have the desire to learn that particular curve at this moment, sadly).
After some research, I’ve found recommendations to use other terminal libraries, or terminal emulators, written in pure Go or supporting Go bindings. I also did some further digging into the roguelike “genre” (I use that term very loosely here) and development process and realized something pretty major. I don’t want to just follow a tutorial and make something that tons of other people are cranking out. I want to make something new. It doesn’t have to be good or exciting or popular. But I want it to be new, and I want to do it right. I’m still set on doing it in Go, but I’m more than likely going to be writing my own generation, pathfinding, and FOV algorithms. It’s a bit more than I initially expected to undertake but I’m ok with that. I’ll still follow the tutorial but only as a checklist, really, to ensure I spend time creating all the same things I would have if I used the libtcod library.
This series may stretch much longer now and it may take me a long time to finish. But I’m going to stick with it until I can release a finished product. I’m not going to guarantee that each post between now and that future completion date will be about Erebor because I’ve got a marriage to maintain, and two kids to raise. Sometimes, I won’t have a lot of time to dedicate to the project. But I will continue to chip away at it in whatever way I can and post meaningful updates as I do.