Next Game

I’ve always been a native JavaScript developer. I rarely force myself to use other programming languages to solve problems.

So ever since I started making games, I’ve been working with JavaScript engines.

In the past, I used many game engines developed by Chinese companies—mainly because HTML5 games were a big trend for a while. But once the hype faded, most of those engines got stuck in that era. Later, WeChat mini-games gave those companies a temporary second wind.

But honestly, they were all terrible. The kind of terrible where you can feel they were made by a team of bad developers.

Cocos Creator was terrible too. No better or worse than the others. But it had one key advantage: a very active community. Back then, before AI chatbots, developers were helping each other out on the forums every single day.

I did consider switching to Unity and using C# to build something more polished. But I couldn’t step out of the Chinese-speaking world—I already had a small following on Chinese social platforms, and I had to develop games quickly using the skills I already had.

In the end, I never achieved what most people would consider success. Publishing and operating a game in China is incredibly difficult.

Looking forward, I’ve set a few game development goals for my personal free time:

1. A 2D Dogfight Game Made with Unity This was actually the second game I ever worked on (back in 2017). Lacking both technical skills and creativity, I somehow gained a bit of attention online. Unfortunately, I built the game using a puzzle-game-style gameplay model, which made everything look and feel off. The development went off the rails, and both the story and visuals turned into a mess.

I thought about rebooting it, but the original gameplay model definitely needs to be thrown out. I considered reimagining it as a roguelite in the style of Vampire Survivors, but after just one day I discovered someone had already made a roguelite + shooter hybrid—and it wasn't even very fun.

I may borrow some elements from Soulslike games, and draw inspiration from old Treasure-developed shooters to rebuild it from scratch.

2. An Epic Game Built with Unity This is one of my life goals—possibly the last game I’ll ever make. But I won’t start until I have enough experience.

3. An Online Social RPG You’ve seen these kinds of games before: you get your own room, you can decorate it, you make friends, and they can visit your room.

In my imagined Meow Meow Land Universe, there's a luxury cruise ship that only children can board.

To gain experience for Game #2, I’d like to develop this one first.

The problem is: this game requires online servers. Without them, it’s nothing—it can’t have an offline mode.

What I didn’t know before is that services like Cloudflare Workers (a serverless platform) can handle long-lived WebSocket connections. With Durable Objects, I can maintain stable online services for just $5/month.

As an engineering-oriented developer, I want it to be robust. The engine it connects to matters.

I’m not sure Unity can handle this well—but I’m certain a native JavaScript project will do better.

I probably won’t use Cocos Creator anymore. I’m not sure Phaser is good enough, but I’m going to try it.

A Final Note on Packaging I need a stable, cross-platform build solution.

Cocos Creator’s build system is a nightmare. Even today, they still haven’t officially explained how to build a Windows executable.

Games made with Cocos Creator usually run fine in the browser, but once you try to connect it to the underlying Cocos2d-x layer in weird ways, everything falls apart: horrible performance and endless platform-specific bugs.

Luckily, I discovered Tauri. I hadn’t checked on it in a while—I used to think it couldn’t package for Android or iOS.

But after trying Tauri v2, I was blown away. It’s fast, stable, and incredibly performant. I have no idea how they pulled it off.

And once I have Tauri, I don’t need to care about how any specific JS engine handles publishing. I’ll just throw the build output into Tauri and export executables for every platform in one click.

Simply put, I’m really confident in this upgraded stack.

Phaser + Tauri + Cloudflare, and once more—let’s make a game in JavaScript!