Hacker News Reader

Comment by mewse | Hacker News Reader

Geez, now I want to write an article about how a game engine works, explaining the rationale for every piece of it; I don't think I've ever seen somebody do that before. It's all kind of spread around in disparate bits and pieces, often quite hard to find (or even to know that you ought to be looking for it).

Problem is that such a work would probably have to be more of a textbook, in terms of length and impenetrability. And it's not immediately obvious who the audience would be, as the vast majority of game developers these days are using pre-existing engines, rather than building new ones. But maybe there's still enough conceptual underpinnings that are worth discussing which would still be relevant for Unity/Unreal/Godot folks.

View on HN

I think there is a middle ground. You can give a bird's eye view of the way things are done and a hand-wavy intuition for why they are done that way. Analogies go a long way towards building the latter. Then for those who want to dive deep, you can leave pointers to articles that cover such things in detail.

It doesn't always have to be a choice between writing a book and nothing at all because the topic is too complex. I personally have benefited from bird's eye view blog posts and mini-articles more than I can remember. If you have the knowledge then go for it, someone out there will thank you.

I think with most posts and also textbooks the problem is scope. I think the Handmade Hero project shows quite well what an amount of work it is to explain it. And there is the difference between a generic game engine like Unity and a specific engine for a specific project. And I mainly mean platform targeting and the likes. But I would read a blog post series :)

I'm sure there's way too much variability to write a whole game engine and describe every detail as "this is the best way to do it because X" rather than "this is the way we did it because personal opinion/time/money/we had no clue how it works."

When it comes to the 3d renderer alone, there's at least enough common ground to explain the big ideas. Like "we multiply matrices because GPUs can do that very well in parallel and we describe our geometry as points because we can then describe the whole image using simple coordinate transformations which boil down to matrix multiplications."

Yeah, I agree that you'd never get anywhere with a "this is the best way to do it because X" approach; it's just too divisive and often there's no single "best" answer that can be equally applied to every project.

I feel like a comprehensive article would need to be more of a catalog of "here are the basics of the various major approaches and how they differ and why you might choose one over another", rather than (for example) advocating specifically for forward shading vs. deferred vs. forward+ vs. clustered as being the universally "best" option. (especially when that whole discussion has become kind of moot for the majority of modern games developers, who mostly just take their engine's default rendering pipeline and just use it)

Reply on news.ycombinator.com