FAQ
Frequently Asked Forcibly Anticipated Questions
- Who are you?
- What are you building?
- Will you be open-sourcing your work?
- Can you tell me more about the engine?
- Do you have NIH syndrome?
- How was this site built?
- Do you use “AI”?
- How can I support R&P?
Who are you?
Rook & Possum is a two-person studio, working on a number of things including:
- A game
- An engine for said game
- Learning materials for people interested in doing either of the above
"Scuff3D" Rook is the solo developer who is working on all things tech. Rendering, physics, audio, animation, systems, networking, scripting, and so on. He was formerly a principal level engineer at various studios, focusing largely on graphics and low-level systems.
Possum "the Opossum" is the artiste who is focusing on 3D content creation, UI, FX, animation, and various aspects of game design.
Unlike a traditional studio, one of the principal goals of Rook & Possum is to give back to the game dev community via educational content.
What are you building?
There isn't much we can say about the game yet, although the game admittedly not the immediate focus. For now, we're establishing our foundation.
We are intentionally not using commercial engines, or even existing open source engines for this endeavor. One of the primary goals is to build everything needed to ship the game from the ground up, and document the process for the community.
This is obviously not the "fastest route to a game." That's not the goal.
Will you be open-sourcing your work?
Not at this time. We have nothing against open source (quite the contrary), but shipping and maintaining open source libraries for the community would add an additional burden we aren't able to accommodate.
The focus for R&P will be to release content (text, video) that explains the design process, architecture, and main insights needed to build what we're building. Distilling this information down to an interested audience requires quite a bit of thought already, and in our opinion, this is a far more valuable offering we can provide.
If we choose to open-source anything in the future, it will be because we are willing to commit the time to maintain it.
Can you tell me more about the engine?
The engine is referred to as the Rookery, and its built on years of experience with various AAA engines, both in-house, and commercially available ones.
The Rookery is written primarily in C++, so some posts may be specifically C++ oriented. However, the majority of the content here is expected to generalize to any programming language environment. Performance is an area where this engine seeks to be uncompromising wherever possible.
Premature optimization is the root of all evil, and I am evil incarnate.Scuff3D Rook
More seriously though, quite a bit of effort will be made to make Rookery systems fast and production ready, even beyond the capabilities of what we need. Setting the bar higher than needed is sure to create learning opportunities for us, and for hopefully for you! The other thing to consider is that if we commit the grave sin of premature optimization and find that the optimization wasn't worth it, that in and of itself is a valuable lesson. Too much oxygen is wasted talking about why something shouldn't be done as opposed to just trying it and pivoting if it turns out to be a bad idea.
The engine has only had a few months of development at the time of this writing, but to get a sense of the scope of the ambition, here are some facilities implemented in the engine currently (as of May 2026):
- Custom build system, referred to as the Hatchery, capable of building not just the engine, but also itself, shaders, textures, fonts, and other assets.
- Custom memory allocator (similar performance characteristics to mimalloc and rpmalloc but with several key architectural departures).
- Custom data structures supporting zero-copy memory-mapped serialization.
- Math library with design decisions geared towards easy CPU and GPU interoperability.
- Multi-threaded DX12 renderer backend.
- Coroutine based task system.
This is obviously still very far from being an engine that can build a game! As the engine is being built, the blog will be fleshed out with more posts describing the various systems. Still to come are things like animation, UI layout engine, materials, lighting, physics, terrain, and so on. Initially, more of the content will be text based, but we hope to transition to more video as the engine gains additional capabilities. You might argue that this content is being released too early, before stuff is really ready for primetime. In our case, we're consciously trying to show as much work-in-progress stuff as we can and publishing consistently is what we're aiming for.
Do you have NIH syndrome?
Because R&P is focusing as much on developing educational content as it is the game itself, there is a natural bias towards "rolling our own" instead of relying on existing solutions.
Building things with reckless abandon is really the only way to truly understand how something works. Furthermore, the exploration can sometimes yield surprising opportunities that may be missed when being strictly content with the status quo.
In a traditional software environment with harsh deadlines on large teams, in-house solutions can admittedly be more of a liability than an asset. In the case of R&P, if something is tried, and it doesn't work, that can still be a good outcome if we can communicate what went wrong for the benefit of the community. Failure is only a failure in the absence of a lesson well-learned.
How was this site built?
This site is built offline using a static site generator. It has zero tracking code, and makes no server requests aside from the requests needed for the actual content you are reading. The layout was designed from the ground up in vanilla HTML and CSS, with the goal of achieving fast render times and high legibility with both dark and light color schemes. The layout was tested with increased browser view scales for accessibility.
Math is rendered using KaTeX offline to ensure the page renders as quickly as possible when viewed.
Site colors are taken from the lovely Catppuccin palettes. Icons are taken from the Phosphor Icons set. The site uses two fonts, Geist, and Geist Mono as complementary fonts for displaying prose and code.
The site functions without JS, but JS is minimally used in two locations. First, JS is used to allow the user to toggle the color theme between a light and dark scheme and persist the preference in local storage. By default, the scheme chosen is based on the user's system preference. JS is also used to track local scroll state to highlight the active section in the table of contents of a post. On mobile, or any situations where the table of contents isn't visible, this code deactivates to reduce the carbon footprint of this site as much as possible.
JS may be used in the future to power local web demos or visualizations. Any post that does so will note as such near the top, and all demos will be set up to deactivate when not visible on screen.
Do you use “AI”?
If by “AI”, what is meant is Generative AI or LLMs, we have made the personal decision to use neither in any capacity.
On the art side, we do not believe the generative AI models today were trained on content under the spirit of fair-use guidelines. Even though the court of law may not have clarified this in writing, this is just the opinion we hold.
On the code side (and math and learning in general), the utility provided by LLMs undermines my (Rook's) own understanding, which inhibits my ability to communicate effectively about any given topic. Wrestling with a hard problem, and not having immediate results is an important part of the learning process. Using LLMs feels tantamount to trying to learn from a math book with the answer key wide open (not recommended).
On the writing side, everything you read on this site was written and edited by a human. We won't shy away from offering our opinions on things, and sometimes, we will be outright wrong. We'll have typos, and mistakes, so don't expect "sterile writing." However, we will make an effort to amend mistakes as we discover them. This way, what you see, watch, read, and hear is uniquely our voice.
"AI" when mentioned on any other page on this site will refer to the traditional usage of the term AI in the gamedev sense to mean systems that govern autonomous entities in the game simulation.
How can I support R&P?
First, we're very flattered that you'd even consider this. All materials we intend to post here will be available to everyone, even if no direct support is provided.
The main way you can help is by reading our content, watching our videos, learning with us, and providing feedback. If you benefitted from something we did, let us know, and let others know too!
Last, if you are in a position to contribute financially, that is of course appreciated as well. The two of us are operating fully independently, without any outside investment and we intend on remaining independent indefinitely.
Each post and video will have a corresponding post on Patreon. Members that decide to support us will be able to comment to leave feedback or ask questions. Occasionally, we may also publish posts that give a preview of future content and solicit feedback from the community.