post-jam changelogs & thoughts

the plan for updating this game was to optimize, polish, and expand. I ended up optimizing and polishing a lot, but I didn’t really expand the content — all the levels have been reworked, but the total amount of content is about the same. a full playthrough takes a few minutes.

I have a lot of ideas to expand on this concept, but there are some technical challenges I would need to solve first — especially SDF-based physics collisions. for now, I’m stashing all my ideas for a possible “sequel” (no promises) while I work on other projects.

post-jam changelogs

v1.1.0 — 2024-04-23

% improved performance
+ expanded graphics options
+ added two new moves (walljump and quickstop)
% tweaked movement physics
% reworked all the levels
+ added a “proper” ending
% fixed visual tearing (hopefully)
% fixed audio crackling
% fixed seams in geometry
:( unfortunately introduced shadow banding in some areas
:) a bunch more polish and tweaks

v1.1.1 — 2024-04-24

% adjusted the options menu
% modified some levels
% polished the ending

v1.1.2 — 2024-05-04

% reworked walljump (less jumpy, more climby)
% modified some levels (again)
:) removed most of the shadow banding introduced in v1.1.0

jam retrospective

my favorite part of this project is the movement physics, I just really like how it feels. also, this was my first time making sfx and writing raymarching shaders, and I’m really happy with how it all came together during the 11 days I had for the jam. I only wish I had the opportunity to develop more gameplay mechanics and make more levels, which is becoming a recurring theme for me in jams. oh, and it’d be nice if the game didn’t require a supercomputer to run, but the post-jam updates will hopefully improve things a bit.

the future

with Geometric Dreams, I’ve bare scratched the surface of what I wanted to do with this idea. however, there are some major roadblocks that necessitate a lot more work than I can put in right now, so I’m postponing all my ideas for a future “sequel” — though I can’t promise it’ll ever happen.

the main problem is the lack of SDF-based physics collisions to pair with the raymarched visuals. currently, the game simply iterates through CSG nodes in the scene, and compiles a shader to render that geometry. it’s really tricky to do anything interesting with the level mechanics when everything has to be synchronized between CPU physics and GPU raymarching. and all the cool raymarching features like repetition and warping become practically infeasible with traditional approaches to physics. also, performance is dependent on the amount of shapes, which is just another hurdle for the level design. another glaring issue is the skybox, especially the harsh seam it creates at the horizon, but I won’t get into that.

I think I would need to hire a raymarching shader wizard to actually pull off what I want to do with this idea. but for now, Geometric Dreams is on hold while I explore other ideas.

keep dreaming!

— kasdf