YAVE began as an exercise in order to hone my skills in graphics and game programming, but I have become confident enough in its potential to keep working on it.
For YAVE, I am focusing on tech features that I believe will make for more engaging presentation in next-gen voxel games. Here are some of my goals (not promises):
- Very large game worlds (> 2000 voxels tall) with very fast movement speeds (> 100 voxels per second)
- Average-case optimizations such as Occlusion Culling and terrain LOD
- An emphasis on optimization for high end machines, especially those with SSDs
- Lights with a large radius of effect (think Stadium lights)
- Fancier lighting in indoor environments
- 3D sound
- More handcrafted world generation, at the cost of generating the world all at once.
- One top secret feature which I have prototyped, but which won't come to YAVE for quite some time
Conversely, there are some features which YAVE will not have:
- Small or "smooth" voxel terrain
- "Science sandbox" features like planets, infinite generation, erosion simulation, etc.
Finally, I believe that voxel games should have an art style embraces the restricted "canvas" that a voxel game presents, and doesn't try to give a realistic presentation to fundamentally unreal worlds. Therefore, instead of looking like this, I'm hoping that my demo for YAVE will emulate the look of artistically confident low-poly games from the PS1 and PS2 eras, such as Mega Man Legends or Wind Waker.
I believe this set of goals distinguishes YAVE from the other voxel game projects out there. Voxels have a incredibly wide range of gameplay applications, so I believe there's enough room out there for a few more voxel games. Lastly, there's one specific and crucial difference between OBBG and YAVE, and that is that I am significantly less experienced than Mr. Barrett.
From time to time I hope to hold streams. Rather than a fully covering the project's development, they will serve as progress reports and code reviews. During streams, I can discuss problems I'm working on, and viewers can suggest improvements to the code.