After university, my first job was as a software engineer at Rare. I joined its R&D team and worked on what became known as REngine: a shared 3D engine intended for several future GameCube projects—though not Perfect Dark Zero, whose team developed its own technology.

During that time, I began designing the 3D engine I would build without the GameCube’s technical constraints. The ideas included tessellated surfaces, spline-based animation, soft-body skinning, and “correct” collision detection.

I’ve been exploring parts of that design ever since, through OpenGL experiments and general-purpose CPU soft-body solvers. More recently, I began working with Vulkan, got the foundations of an engine running, and started combining my various projects into a single repository: fireEngine.

While bringing those experiments together—with AI-assisted testing and some fairly fierce reviews—I realised that I wasn’t happy with the overall architecture. I wanted to rebuild several areas around a clearer and more structured design. For example, I would like the rendering subsystem to use a render graph similar to the directed acyclic graph (DAG) used by Maya.

Rebuilding the engine from the ground up, while documenting the decisions behind it, seemed useful both for me and for anyone interested in learning about the techniques involved.

With that in mind, I created fireEngine-tutorial: a step-by-step guide to building a 3D engine from scratch and restructuring fireEngine around a more coherent architecture.

Each repository release will have a corresponding blog post explaining what changed, how the code works, and why particular design decisions were made. The repository will hold the step-by-step history of the engine, while this site provides the companion narrative: the rendering techniques, experiments, dead ends, trade-offs, and lessons that do not fit neatly into a changelog.

Follow the project

• Read the latest posts.
• Browse the fireEngine tutorial source.
• Subscribe to the Atom feed.