Qubi Research Program Postmortem

Published on 2021-06-26

Table of Contents

Hello world,

After a long 8 week hiatus, I’m finally back with a new game on itch.io.

It’s called “Qubi Research Program” which is a rather “unique” typing game. I came up with the idea for it after reading the theme for the 83rd Mini Jam which was “Dread”.

With the word “Dread” in mind, I immediately thought about a shitty office job which one require in order to survive. Wouldn’t it be silly to have a game which replicate Stanley’s job but have the dreadful feeling of living within a broken system as Papers, Please.

Proof of concept

During this time, I was learning Nim and Raylib so I decided to quickly make a proof-of-concept.

The proof-of-concept is only the typing game part, just to see how it feels but the idea is there.

At the end of each work day, the player is awarded with a compensation based on the amount of words you’ve written. Each typo made are fined.

The game grows difficult each day as you have to pay various bills and make sure that your account isn’t surpassing the fixed overdraft limit.

My close friends liked the idea and told me that I should make it even if it couldn’t submit it for the 83rd Mini Jam.

The game uses the same common English words dictionary as Up Goer Five and Thing Explainer. The reason is simply that I want to only use words which everyone have heard and used before.

Initial foothold

I started writing the game on Friday, 25 June 2021. I decided to do it on a public repository in my personal GitLab instance because why not.

There’s not a lot of games written in Nim surprisingly so I think it could serve as a relatively complete example of a simple game.

I started by writing a simple Hello World and a Nix shell environment which have Nim and Emscripten.

Emscripten / WASM has proven to be quite annoying at times. One of the annoyances is the cooperative multitasking model which cause the entire program to freeze if it is shoved to the background. I’m not sure how to fix this bug and couldn’t really find any good references on how to fix it.

However, once I have a working toolchain setup and make it possible to build within Nix imperatively, it goes smooth sailing.

Constructing the Foundation

The architecture of the game is very simple. It uses screens which encapsulates a state along with methods for screen initialisation, update and deinitialisation. Each screen are simply a subclass of the Screen abstract class. No fancy ECS required for something simple and quick.

While I am not a fan of object-oriented programming, something like this shows why OOP can be good sometimes especially when you need to encapsulate both a state and an interface.

I start to appreciate Nim’s approach in making both functional / imperative / object-oriented approach work together nicely.

In particular, I like that f(x) is the same as x.f. This is similar to the pipeline operator which are popular in functional programming languages. This makes f(g(h(x))) readable by transforming it to x.h.g.f.

Of course, there are a lot of things inside Nim which I love: Portable compiler, Sum types/ADTs, Macros, Syntax clarity, Infix functions, etc.

Honestly, I wish more programming languages reflects Haskell’s syntax clarity. I hate those which have a lot of noise such as Rust. However, I think I’ll use Nim for things which does not make sense to do in Haskell.

Basic Gameplay

Porting the code from the proof-of-concept to the foundation is fairly simple. Most of it just moving the stateful variables within the encapsulated state of the GameScreen class.

Once that’s done, I just started working on the “Day End Screen” and the “Game Over Screen”. The game progress is saved outside of the encapsulated state. In other words, they’re just global variables.

While a lot of people cringe at mutable global variables, it’s fine for something quick and dirty and who cares, really. I’m not writing software, I’m writing games.

After I got the basic gameplay / minimum viable product, I immediately share it with close friends.


Originally, I planned to have a day to be a minute but during initial playtesting, I deemed it to be too long so I cut it in half. a week was 5 days but it is also too long, I end up cutting it down to 3 days.

So in this game, a day is 30 seconds and a week is 3 days. You also pay rent every week which increases every week too. I would like for the game to also last a bit long but not too long, unless you really want to.

You get 1 point for each word done but you’re fined 2 points for each error you made. While this seem excessive, this is deliberate to make it difficult and have that “dreadful” feeling.

The initial version was fairly difficult even for accurate/fast typists. No one was able to surpass 15 days / 7 minutes of gameplay.

To balance, I start by laying down some ground rules. I think it’s unreasonable to have a speed of 1 word a second. So, I cut it down by 20% in order to have a bit more room for errors. In the real world, rent is ideally at most 60% of your income. With the daily goal value in place, I simply make the initial rent to be 60% of the weekly goal.

In the real world, 10% rent increase is excessive but this is a game and games need to be short so 10% it is.

With the balanced changes in place, it is now easy to hit more than 30 days. Nice, now I have more room to make the game more difficult

Increasing Difficulty

The first thing I did is to have the duration of each day randomized to destroy the 30 second rhythm.

Second, I want to shuffle the characters in a word which reflects Typoglycemia / Transposed letter effect. This is to make it even more annoying once you have the rhythm and expectation to write English words.

Third, like in “Papers, Please”, I want to have “surprise costs”. Paying for random things such as coffee, parking, and student loans.

The factor of each “surprise cost” is simply a prime number from 2 to 59. At each the end of the work day, A random value from 2 to 100 is generated and if it’s multiple of a suprise cost’s factor, the surprise cost is added. I’m too lazy to do something proper and this will do just fine.


I think I’m done with this game. It’s fairly short and simple and it is more of an excercise for me in learning Nim and Raylib.

I love both Nim and Raylib and I think I have found a good stack to write small games with. I hope with all of this that I will start creating games a bit more frequently. No promises though :-)

Thanks for reading this rather messy and all-over-the-place of a postmortem. As of writing, it is 5 AM in the morning and I’ve been awake since 12 PM. I should go to bed…

The game is available on itch.io and its source code (ISC) is available on GitHub.