Blog

Another year went by as an indie game developer and what do I have to show for it?

In last year's retrospective I wrote that apart from working on my game The Big Forest in general, I had four concrete goals for 2024:

• Present my Fractal Dithering technique
• Release my Layer-Based ProcGen for Infinite Worlds framework as open source
• Wrap up and release The Cluster as a free experimental game
• Make better use of my YouTube channel

I ended up doing only two of those, but it was the two most important ones to me, so I'm feeling all right with that.

Release of LayerProcGen as open source

I released my LayerProcGen framework as open source in May 2024. LayerProcGen is a framework that can be used to implement layer-based procedural generation that's infinite, deterministic and contextual.

I wrote extensive documentation describing not only the specifics of how to use it, but also the overarching ideas and principles it's based on. I also did a talk at Everything Procedural Conference about it, which was well received.

I'm unsure how many people are using the framework directly for their games, but I know of several who've been inspired by its underlying ideas and made their own implementations.

Sythelux Rikd ported LayerProcGen to Godot. The core framework worked in Godot out of the box, but Sythelux ported various of its Unity-dependent optional utilities and made example scenes for Godot. Oli Scherer created a Rust implementation of LayerProcgen with some slightly different architectural choices. For his game "Around The World" Thomas ten Cate aka Frozen Fractal wrote his own implementation that works on a sphere.

Overall I feel like LayerProcGen has made at least a little dent in the procedural generation community, and I'm happy to have been able to make such a contribution.

Release of The Cluster as a free experimental game

I finally wrapped up and released The Cluster, a game I'd been working on on-and-off since 2003! I already wrote about it here, so no need to repeat myself further, but despite not having much of an impact on anything, I'm personally happy to at long last have this "unfinished business" out of the way.

Stuff I didn't get to do

I didn't get around to presenting and releasing my Fractal Dithering technique. Hopefully this year!

I also never made more "proper" YouTube videos after the one in January 2024 about the terrain generation. I considered doing one about LayerProcGen, but there's already the video of my conference talk about it. I could probably explain the framework and ideas slightly better by creating a video with nice animated graphics and diagrams, but it would be a lot of work for marginal added utility. I also considered doing one about all the interesting tech in The Cluster, but I didn't feel like diverting more of my time to that game over working on my current game. Why no new videos about my current game The Big Forest then? Well, read on...

Progress on The Big Forest

The Big Forest is the game I'm currently working on, and it's still in its early stages, just like I said last year. The game will be fully procedurally generated, and so far I've been working on a series of disconnected experiments and proofs of concept that will eventually all be rolled into the game. These include procedural generation of terrain, gameplay progression, creatures and music.

I had worked on the terrain generation in 2023, and I continued this work in the first few months of 2024.

• Previously the paths in the terrain had simply gone from one chunk corner to the diagonally opposite one. I now generated points of interest (with towers for now) and made the paths connect those, although it's not yet completely robust.
• I managed to radically speed up the generation by making use of Unity's Burst compiler (without Unity Jobs, which are not a good fit). I'm very happy with that. For fun, I made a video where I had drastically increased the player's speed, like they're some kind of fast hedgehog.
• I changed the terrain chunk implementation to use chunks of different sizes, with larger chunks of lower resolution further away from the player. This made it possible to expand the draw distance greatly without slowing the generation down.
• Using chunks of different resolutions introduced cracks in the terrain, so I implemented a skirt solution.

On the subject of YouTube videos though, I'm not sure I could make an interesting video of the 2024 terrain progress mentioned above, and there's plenty of other videos out there that cover very similar ground.

After this I kind of needed progress with other aspects of the game before it made sense to focus more on the terrain. After taking a break to wrap up and release LayerProcGen and The Cluster, I returned to The Big Forest to work on its procedural creatures.

I spend the latter half of 2024 working on that - three months on creature model generation and three months on procedural animation. I was writing a whole section about that here, but it got so long that I've written a whole separate post about the procedural creatures instead. Long story short though, I had some progress but it's still very far from where it needs to be.

The fact that it's still well short of being any good means I don't consider it good material for a YouTube video either, as I think such videos work best when they can conclude in a satisfying wrap-up of something. I did write the blog post though, and now I need a bit of a break before I resume working on the creatures again.

Oh right, a video of a bug in the procedural animation I was working on became my most viral post ever on Twitter, Bluesky and Mastodon. (The creatures in this video are not procedurally generated; I was testing the procedural animation on purchased animal models for research purposes.)

Other tidbits

I can take no credit for it, but my business partner for the Quest port of Eye of the Temple, Salmi Games, released an update for Quest 3 that adds real-time shadows from the torch, higher frame rates (90fps almost everywhere), improved texture resolutions and aniso, and higher clarity due to foveated rendering being mostly disabled. This brings the Quest version (on Quest 3) even closer to the PCVR version. UploadVR covered it here.

I very occasionally make music, and in 2024 I released two very different tracks. And as a first, these ones even have vocals (by me): Just to Say Goodbye, a sad song about an alien with bad timing, and Sorte Gryde, a metal re-imagining of an old Danish singing game, with a dash of Saturday-morning cartoon villainy.

I was briefly mentioned in this video about making games feel mysterious by Mark Brown of Game Maker's Toolkit due to my 2021 article about designing for a sense of mystery and wonder. Neat!

When I was around 11 I would draw "video games" on paper during breaks at school. Not sketches or design docs, but games that were fully playable using your finger to indicate where your avatar is, and following simple rules. Earlier this year I had these games scanned (42 pages!). I made a zoomable annotated display of one of the games. I'd like to do more with the other games too, maybe even make a (digitally) playable game out of the one with 3D perspective dungeons.

Goals for 2025

I'll keep this short.

• Develop the creature generation and animation sufficiently to put them into the game. If I'm still not there by 2026, I need to seriously reevaluate some things.
• Make at least one new YouTube video about The Big Forest.
• Present my Fractal Dithering technique.

Wish me luck!