tinywars - WTF, I'm making a game?

August 13, 2021

Bitmap Brothers ❤️

tinywars - WTF, I'm making a game?

Why, hello there. Long time no see. You may know me from such failed projects like Wee or Qak. Today, I'm here to tell you about my next big disappointment: tinywars!

A few weeks ago, Kev of Coke And Code fame, twerterered about his latest game, Tales of Yore. It's a minimalist, WIP MMORPG and it just blew me away. Go check it out.

Kev's applying a very iterative "I'm a dad and can only spend a few hours a week tops on this" development mode. And despite these somewhat rigid constraints, he managed to create a great little gem which gets better and better.

In light of the arrival of our newborn son 4 weeks ago, that seems like a great activity in between diaper changes, feeding sessions, and the especially creative baby entertainment sessions, consisting of fun activities like "bamboozle the baby with differently colored book backs on the bookshelf" or "bamboozle the baby by badly drawing on a ceramic pig with chalk".

Draw me like a french pig.

While Kev went straight to "Science-based MMORGP with dragons" for his passion project, my ambitions are a little more modest. There was one game in my childhood that really left an impression: Z, by the glorious Bitmap Brothers. Give it a try in the embed below!

Z was different from other real-time strategy games at the time. Instead of base building, you conquer sections on the map by capturing the section's flag with a unit. Once you control a section, you can use the section's production facility to build additional units. If the section is taken over by the enemy, the facility is immediately theirs to use for production.

A similar mechanic applies to vehicles, which can be found around the map, or build in production facilities.

Almost everything on the map can be destroyed, from rock formations, to bridges, and even production facilities.

As the game progresses, your units gain experience, which translates into better aim and evasive maneuvers when attacked. You kind of get attached to the little fuckers.

Each party starts out with one section under their control, which also houses their fort. The game ends when the forts of all enemy parties have been destroyed.

All of this made for comparatively fast paced, tug-of-war game play. No time was wasted on building a beautiful base. It was all just "grab section flag", "build units", "pew pew pew".

I've never played Z in multi-player mode, but I always had a feeling it would lend itself incredibly well for online play.

So that's what I want to build. A Z inspired multi-player real-time strategy game. No base building, just "pew, pew". I'll leave the detailed game design for later. I think it will naturally emerge from the basic mechanics I want to implement bit by bit, as time and baby permits. Accident driven development.

Stack-edy Stack


If you guessed I'd be using libGDX for this, you'd be wrong. See, libGDX is fine. But I also want to use this opportunity to learn. And complain. As I've complained enough about libGDX in my life, and since this is going to be a browser game, I might as well learn all the fancy tech them young whippersnappers created over the past 10 years!

I'll be using TypeScript for both the front-end and back-end. TypeScript's been with me for a while, and I think it's the best thing that has happened to the web in the past years. Thanks, Microsoft. I guess?

Instead of reinventing the wheel in terms of basic game loop, input, rendering and sound, I'll be using Kev's brand new Gute library.

And I guess that's where the fun ends and the tortu^H^H^H^H learning experience begins.

The server side will be implemented using Node.js in combination with socket.io as a thin wrapper around Websockets. I'm unsure yet whether I should serve static files via Nginx or do it via Express from within the Node server.

For the front-end chrome, like lobby and chat system, I'll likely make do with just vanilla JavaScript plus the DOM. If I feel adventurous, I may give Vue.js a try.

Daddy likes short iteration loops, so the biggest hurdle will be figuring out a good development workflow for all of the above, including such newfangled things as IDE support (VS Code), debugging, and deployment.

Since this is a networked real-time strategy game, there'll be plenty of opportunity to explore topics like networked lockstep simulations, path finding, group movement, and basic RTS AI (in case I add a single player mode).

The plan

How I see myself in the future.

I'll just wing it. Which means I'll get the boring bits like the development workflow and deployment out of the way, and then work out basic game infrastructure and mechanics. Once that's done, I'll have to tie it all together into an actual game. I've written an RTS before so there's a good chance I might actually finish this for once.

I'll need help with graphics and audio and will try my luck outsourcing that for $$$ on Twitter. I may also team up with someone who can create a coherent story for the world the game will be set in and possibly help with game design-y things and balancing.

I'll document my findings along the way on this very blog. I will likely curse a lot, both IRL and here. And maybe some useful text articles on game dev-y topics may fall out at the end. The previously failed projects I documented here mostly failed because writing about them became a burden. We'll see how this works out this time. I might just prioritize actually work on the game.

Up next

Next time I'll take you on a wonderful journey of figuring out what the fuck a Node.js is and how to best stab myself in the eye with it.

Discuss this post on Twitter.