It used to be that when I started a new RPG I got this amazing feeling of excitement when I began to explore the world. The thrilling feeling of entering a new area and wondering what secrets could be hidden, waiting to be discovered.

Getting older, I experienced this less and less. The novelty of exploring a virtual world seemed to wear off as I got older. Then in 2011 I (along with the rest of the world) found Minecraft. The experience was fresh and exciting. A great implementation of a different approach to generating a virtual world. Something about knowing that world is infinite and unlike the world that anyone else is playing made exploration feel meaningful again.

Once I began to internalize the patterns of terrain in Minecraft the feeling of excitement began to ebb. It was still exciting to find diamonds, but the random terrain began to feel repetitive and similar. I wanted to find something not random in the world - I wanted to find something that felt intelligent. What if there were a way to stumble on the ruins of other users civilizations, decayed after many years of abandonment?

Could a hybrid approach to world building be the next step in the evolution of wanderlust in games? I started writing this game for mobile devices, but my skills at the time were not enough to implement the networking required to simulate the dynamic world I imagined. Then I found Meteor, a javascript framework that makes it very easy to synchronize client data. I re-implemented the game, migrating from the the Moai game engine to Meteor with the front-end entirely written in a custom canvas based game engine. Here is my lightning talk at a 2013 meteor meetup in San Francisco.

There were some pretty awesome features in the last meteor version. Users could write their own custom servers, characters could move between servers with a complex three way authentication handshake. But I wasn’t following Ken Levine’s excellent game development advice: “Stop trying to be smart, and start trying not to be dumb”. The engine and features were too complex, and it didn’t make sense for me to be maintaining my own javascript game engine. So in early 2017 I started re-writing the game again. This time the server is written in Go, and the client built on top of PIXI.js canvas engine. Both decisions have already paid off.

In one way or another I have been working on this game since the first mobile version 2012. It’s nearly always in some part of my mind – even when I am too busy to be actively developing. Whenever I have a spare moment my mind wanders back to the most recent development challenge. There’s some inexplicable force that drives me to keep at it. I envision a dynamic and evolving world in the cloud. A dwarf fortress for the next generation. A Minecraft for the future. A limitless RPG for that can fill others with the same wanderlust that I found in RPGs of my childhood.

Try the current version at aether.world.