Create (or port) a 3D game for Rust

[brson]

Rust should be a great language for multimedia domains like gaming. It would be nice to have a good demonstration of Rust's suitability for high-performance graphics tasks. Servo has a number of Rust graphics libraries that can be useful here, and there are various other useful bindings out there for things like GL and SDL.

[brendanzab]

GLFW3 bindings by yours truely. I'm trying to keep it up to date with the latest updates to rust and glfw3. opengl-es is all sorted via Servo, but bindings for glcorearb.h would be incredibly useful. D has incredible support for OpenGL – it would be great if Rust could be at the same level.

[Subasy]

Hello,

I'm a student from the University of Evry Val d'Essonne, I would like to have more details concerning this project in order to implement it, what the final application should look like, Do you have some example illustrating it please? Waiting for your answer.

Regards, SR.

[Yoric]

@Subasy Well, basically, pick an open-source 3d game of your choice (preferably something rather simple, to begin with), and port it to Rust. You can find a (certainly incomplete) list of open-source games on Wikipedia.

[Subasy]

Is this kind of game : http://en.wikipedia.org/wiki/Frozen_Bubble a valid subject ?

[Yoric]

I suspect @brson wants a 3d game. However, if the resulting game is in 3d, I assume that he will be happy.

[Subasy]

So it means that if the balls for example are created with a library tool instead of an image it will be okay ?

[brendanzab]

I assume that brson wants to be able to showcase Rust's suitability for real-time 3D graphics. So OpenGL would really be what he'd be looking for.

[Subasy]

Indeed I'm not used to code a lot... that's why I'd like to know if you think that the project could be hold by a novice ?

[brendanzab]

Unfortunately 3D graphics programming is very complex stuff. I won't discourage you, as I'm not a very experienced programmer either and I'm working hard learning it, but I do warn you that the road is long and hard. You will experience many days crying in frustration as you attempt to wrap your head around matrices and quaternions.

[Subasy]

Thank you for your help I'll think about it ... but I've just a last question if I'm not missunderstanding I can do something animated using graphics which is not really a video game too ?

[Yoric]

I am sure that this would also be interesting. We will just have to open another issue for this, but that's no big deal.

[Subasy]

I would be interested by this new issue :)

[brson]

Hi, @Subasy.

So this is a very open-ended project, but I have some basic goals for it. The motivation is that multimedia software, such as games, tend to require precise control over memory and CPU resources, benefit from parallelism, and are usually written in tricky low-level languages like C++. Major game developers have been lamenting their programming languages lately. Video game engines seem like a perfect match for Rust, so I would like to prove that out by exercising the language in a way that typical, modern game engines do.

What exactly this means is open to interpretation, but here are some ideas:

• This project is largely about identifying requirements of modern game engines and proving that Rust can be competitive.
• It should use 3D graphics - 3D is more demanding of the language, and OpenGL is a notoriously difficult API. For our purposes, proving the viability of Rust, this is a good thing. Even 2D games regularly use 3D engines now.
• The engine itself (the hard stuff) should be coded in Rust.
• Using Rust for the audio pipeline is also very desirable - audio has very similar processing demands to graphics.
• Demonstrating how Rust's concurrency features can be used to architect a parallel game engine is also important.

The actual scope of the game itself, and whether it is even a 'game' in the traditional sense, doesn't matter so much. The important thing is to demonstrate Rust effectively doing things that game engines do, and hopefully produce new language and infrastructural improvements for future Rust game programmers.

I think that Frozen Bubble is probably not a good example of a modern game engine - it is very simple from an architectural point of view.

[brson]

@Subasy

I do think a project to produce some interesting animation would be great. In particular, I'm thinking a demoscene-style animation, where the focus is on producing attractive results with efficient code. That type of software has a lot in common with video games. Some attractive 3D graphics + simple audio + a parallel architecture = win.

[Subasy]

I think I understand what you're talking about, something like a well designed 3d car rotating with music for instance would be valid ?

[brson]

@Subasy Yes, I think so.

[brendanzab]

As an update, we now have an OpenGL function pointer loader. This is useful for advanced graphics that requires functionality outside of OpenGL ES.