Why I am making a roguelike in 2017

Yes, it is a playground. Yes, it is a pet project. No, it is not viable. Yet, here I am making and building.

The real question is — why would you want to? Have you ever asked yourself that question?

Harsh. If someone earnestly asks me a question about anything I write, I want to respond. This one was hard though. Not really that I take it personally, I have impenetrably thick skin, just that I wasn’t provided any context for the remark.

Though the author probably intended to just troll about, it does hit on the futility of my project. Where is this coming from? Where is this going? Do I have a roadmap?

Yes, I have a roadmap, and this is what 2017 roguelike looks for me:

I have decided on a name: Red Yellow Blue.

I have decided on a platform: native app through Rust.

I have even put together some GUI components.

What I don’t have is a working game, but the dev timeline looks like I will have a version 1 release somewhere around June/July 2017.

This still doesn’t address any of the why question though. Maybe this is all wasted time. Who am I to judge my own work, right?

To explain the backstory, I have spent the last 2–3 months training AIs to play OpenAI games and learning about this deep learning / neural networks stuff. This was fun, and was constructive through the educational aspect. Though, that changed a bit with OpenAI’s latest movements. Specifically, they are moving into NLP, but no open-source to play against. I am not sure what this would even mean to “win” in natural language. Language learning is fundamentally something that only has meaning in relation to some universe context, where there exists objectives to be obtained through use of language.

So, OpenAI just wasn’t fun anymore. I played through the examples and exhausted the space of problems that I could solve on my laptop through modern techniques. Deep learning is incredibly resource intesive. So, a few weeks ago, I had no objective. That is when I decided to fork what OpenAI was trying to do with NLP and just try to do it my own way: and thus RYB was born.

Unlike OpenAI’s approach, my game is a game. In the sense that human players will interact with and hopefully enjoy, playing the thing. So, if that counts as a “why”, then that would be enough for me. For me personally, I don’t really even feel like that is necessary. I have no ambitions for this project other than to have a ton of fun and share that.

Hell, maybe I’ll port it to AR headsets sometime. Progress is slow, but play is constant. The creative energy for computing used to come from just the thrill of making something great. Now I guess everything needs a manager to make you hate your life. Thankfully I don’t have a manager or reporting structure, so I snub my nose at the concept.

I am doing this because because. Long live Red Yellow Blue.