Hello, Pergonauts! Welcome to another Tuesday dev blog for Dandelion Void.
This week we’re offering a work in progress peek into our character creation system, as well as discussing our philosophy of iterative game design.
Celestial Bodies
Player created characters are common in survival games, and with our focus on user customization we are no exception! We’ve actually had a modular character system for a while, but this week is the first time it’s been wired up to the main build.
When testing Dandelion Void, developers now spawn as one of these six presets.
Dandelion Void has an extremely specific aesthetic. Our textures are full of crunchy pixels that don't get bilinear filtering. Our texel resolution is fairly low, which when combined with the isometric perspective, evokes the look of vintage pre-rendered game graphics. Visually, we draw from the mid-century modern and brutalist movements.
Character creators are a balancing act. The player should be able to make a unique avatar that they can identify with, but you also don’t want the range of outputs the tool generates to look out of place against the other game elements. A photorealistic human with anatomically perfect proportions would certainly clash with our lo-fi presentation!
Our solution is to render a few character design elements in an extra stylized way. Notice how while the ribs and muscle definition of the characters are fairly detailed, the eyes are just big blocky rectangles and the body hair (see the legs and chests) is drawn in a checkered lattice? These elements “anchor” the created characters within our core visual motifs, so that no matter what body shape, hairstyle, gender presentation, or skin tone the player chooses, their character always looks like it belongs in this game and in this world.
The next step for this system is to add in a starter wardrobe for these guys, as well as hairstyles beyond “bald” and “mullet.” While the ship launched in the '70s, it's not quite that '70s.
Long-Term investments vs Low Hanging Fruit
It might surprise you to learn that we’ve had this character system for months, but only connected it to the main build this week. This is due to the nature of our production process.
We organize ourselves using a sprint-based “scrum” methodology and a playcentic design process. In simple terms, this means:
We plan our development tasks in units of 2-week “sprints”
At the start of each sprint we talk about the state of the game, and decide the tasks we want to work on.
At the end of each sprint we make a special build of the game and play it together
Our goal is simple: each of these sprint builds should be significantly more fun to play than the one before!
The strength of this system is in the frequent playtesting. Every game developer has had the experience of losing months on a feature that sounded good in their head, only for it to flop the first time someone else played with it enabled. The advantage of testing so frequently is that we get immediate feedback about whether or not something is working, and can change our plan accordingly. This rapidity is why some call this framework “agile.”
At the beginning of a sprint, it's good to start with tasks that represent longer term investments; our junk system described in the last post, for instance, is a unit of new content that took us days to design and implement. Towards the end of the sprint, when we are doing daily group playtests, we pay special attention to easy wins and low hanging fruit; little bugfixes or bits of polish that can be done quickly while giving us a lot of bang for our buck.
Believe it or not, wiring up the characters was one of those low hanging fruits! You see, Brian actually built the backend of this dynamic human system months ago so that we could show multiple characters in the trailer. It even includes support for sliding between different body shapes and weights.
But creating the UI “frontend” for authoring a character – all the menus, sliders, camera angles, etc. – is a whole separate task that we haven't been able to justify putting into a sprint, as our current focus is on improving our core gameplay systems and mechanics.
With that said, we got a little sick of being the same person every time in our playtesting. So we settled on a compromise: the character creation UI can wait for later, but in the meantime, a simple Lua script will spawn the player as one of six random character presets. This was a textbook “easy win”: – all we had to do was wire up a system that already existed, so it just took a couple hours, but it improved our end of sprint build significantly.
We hope this little look into our planning process has been interesting! Please do look forward to a snapshot of the character creation UI in a future post, once it's the right time. We can't wait to see the characters you'll make!