Cube & Star: An Arbitrary Postmortem - Part 1.

I began development of Cube & Star: A Love Story in March of 2013.

I developed it in bursts, sometimes going weeks without touching it - sometimes working for full days on end.

The early alpha won Best “Other” Game in Intel’s LevelUp contest, and was a finalist for Indiecade 2013.

This alpha was uploaded as a “demo” to Steam as part of the former contest.

I usually worked as one half of the duo: "Doppler Interactive" - my partner @TheDopplerDuck is the artist, I am the remainder.

But this time it was a solo project. A lonesome project.

I developed the art, sound, narrative and systems.

This was good and bad. Unchecked cycling over concepts leads to paranoid logic.

The game eventually became Cube & Star: An Arbitrary Love, and was released on Steam for Linux, Mac and PC in February of 2014.

This is the entirely subjective story of its development - in roughly chronological order.

Read Part Two here

Cube & Star: An Arbitrary Love, in its final manifestation.

The Initial Concept

Cube & Star: An Arbitrary Love sprung from an unlikely place: A little demo text-adventure made with a tool I developed for fun, in December, 2012.

That one line in the inventory:

"1 of 6 actions completed".

....combined with the joy of exploration prompted by endless sailing through biomes in Minecraft - combined into one vaguely-experimental concept:

"You are a rounded cube... bouncing through a randomized world. You meet other entities, and you collect things".

Which boils down to the key urges I was attempting to experiment with:

The urge to explore - "What's past that next tree?"
The urge to collect - "I need 10 more coins until I have collected 100%".

These were the core urges that drove my my long-time development partner, @TheDopplerDuck - to play games. Almost trance-like, repetitive but joyful.

I dubbed the concept an interactive homage to her tastes and gaming aesthetic.

The Development

The very first prototype video. You've come a long way, baby.

I began developing Cube & Star in March 2013 - with the intention of working on a "quick" solo project. 

I knew then, and I know doubly now: the "Quick project" is a common illusion.

I was developing with Unity - with the express and arbitrary goal of using predominately procedural content and few if any textures or such.

The Conceptual Mistakes

My key mistakes were made early and were deeply rooted in the intention of the project.

The game wasn't designed for sale or critical attention.
It wasn't designed to excite the market. 
It wasn't designed with any overt rules or goals.
The visual style was a sly jab at the geometric style of the time - but it was too indie. Too simplistic. 

It was a zen little garden with which I could explore some concepts and enjoy the process of development.

There was an emergent beauty in shapes, but in retrospect I feel that the visual style was a core mistake - it didn't translate to a engaging screenshots.

Bringing Color to the World

Initially, the game was intended to have biomes. Water, sand, mud - all represented by single-pixel colors on the tiled map.

The first build featured a random chequerboard of colors and palettes - some beautiful, some ugly.

It had a real candyland vibe.

The urge to explore came through even in this early build.

"What color palette will I find next?" was the driver.

Late at night, I took the world coloring logic and applied it, for fun - to the player - so that they could leave a trail of color wherever they went.

Somehow the feel of bouncing a single unit and stamping color on the ground was fulfilling - in a really small-scale way.

A world without color.

Painting the world.

And so - the vision for the game was revised: Bring color to a greyscale world. The player and the inherent systems would be responsible for creating the biomes.

From here - the core system progressed quickly.

It was conceptually simple:

Entities move around.
When an entity enters a tile, it colors the ground and any scenery in a radius around itself.
If a tile is colored and it has a buried object - spawn that object.

Freedom within Boundaries

The subtleties of the world-creation theory became evident as the builds progressed.

Randomness is good but it is inherently ugly. There is no harmony in randomness. 

I needed to implement randomness within boundaries.

So I set a range of saturation and brightness - within which colors would always be chosen. This enforced a "pastel" style of colors while retaining the dynamic and seemingly random nature of the world.

Pull the Lever, Win a Prize

I'm particularly proud of the way in which collectibles and entities are distributed around the world.

Everytime you pickup a collectible, there is at least one other collectible or object of interest visible on the edge of the screen.

I wanted to trigger an almost trance-like "Just one more horizon, just one more horizon" state in the player.

God rays and collectibles in the evening.

Each collectible had to be rewarding enough that it was worthwhile collecting and trivial enough that it wasn't a core requirement for enjoying the gameplay.

Ultimately - I created a naming system which named items based on color and type.

For example, a "Blushing Diamond" may be an approximately red gem pickup.

All of the words were chosen from lists of synonyms, and the text used to describe your discovery was selected in turn from a list of sentences.

The aim, from a development perspective, was to create massive variety from a small palette of inputs.

The Narrative

The narrative and general message of Cube & Star has received some really interesting interpretations since its release. 

At this stage of development - the game was about non-violent joy. 
The thrill of exploration and color.

To give a tiny bit of structure to the game - I added an "Ancient Cube" - an ominous character who lectures you about collecting stars - the very activity you are about to engage in.

To give a vague endpoint - I added an "Ancient Star" - that one ideal star that the Ancient Cube longs for, from his idealized perspective of his own youth.

Bring color to enough of the world, and the Cube and Star will be reunited.

The initial narrative was fairly transparent. 

Reunite the Ancient Cube with the Ancient Star. 
Leave home and make your own way. 
Defy convention.

Life

I really like the concept of emergence in a game. I like the idea of a system being more-or-less independent of the player. 

The player is just a willing participant in an ecosystem over which they have indirect control.

As biome-creation was delegated to non-player entities, the life of the world was critical.

I added three core variants of entity.

It will take keen eyes - but there's a squirrel, a frog and a few stalkers in there (top right, pyramids).

The Other Cubes were simple characters. They moved one tile at a time, in a random direction. If there was a tree, they would nudge it. If there was a fruit, they would eat it. If there was a star, they would collect it.

The Stalkers were more advanced. They would pick a colorful point on the map and path to it, removing color as they went.

Critters were the most advanced. They would path to a tree and nudge it when they arrived.

The World

The feeling of building something is core to gameplay, I think. 

Whether it's literally building things or building anecdotes it's important to show some sort of change that the player has affected within the game.

In the case of Cube & Star - the world is the canvas, and the player is the most aggressive artist.

When the world is created - it is very simple: A chequerboard of light and dark grey and nine different trees which increase in size as player ventures from the start point.

As the game progresses, both the player and the entities in the world spread color. 

If the color is strong, grass will grow.

If the color spreads to an invisible critter spawn point, a new critter will emerge from the ground.

And through all of this - a time-of-day system will cycle the world through Morning, Daytime, Twilight and Evening.

Twilight in Cube & Star - a few stalkers, a coin and our humble selves.

Juice

Juiciness was core to the experience. The inputs of the player are very limited: Move in four directions. 

I aimed to make every movement rewarding, even if in a minor way.

When the player moves - they will either bounce or roll.

If there is grass underneath them, they will flatten it.

If the ground or scenery can be colored, they will be.

If there is a tree in the way, it will reverberate in a rubbery way, and an item will fall from it.

And if there is an entity in the way, they will say something - drawn from a huge bank of entity-specific dialogue.

Jostling a tree. The life of botanist royalty.

The Vocal Mis-step

I chose a voice synthesizer to give both the Cube and Star a voice. 

I liked the downtrodden and unearthly tone of the technology - and I also wanted to keep the Cube and the Star gender-neutral.

I felt like enforcing a gender on either would severely limit the ability for a player to translate the game in line with their worldview.

This choice would later haunt me, as its grating tone and the low-rent associations of speech synthesis dominated the game's eventual Greenlight feedback.

Next Up

We had established the gameplay and loose concept, we'd constructed a world of sorts, and it's time to move forward with development.

Trees, cubes, stalkers and painting. The "jumps" in functionality in early development are so extreme.

In the next edition of this post-mortem we'll look at the growth of the alpha and the surprising attention it eventually received.

Stay tuned!

-Joshua McGrath

Continue to Part Two here