SPIKE: Multiple renderers vs one renderer and a Renderable class
- 主要语言
- C#
- 星标
- 6
- 派生
- 0
- 平均合并
- 1 小时 1 分钟
- 30 天内合并 PR
- 2
描述
Investigate what should be the best approach in terms of performance, maintanability and use of use for developers.
If all objects can just inherit the Renderable class and tell how a single renderer should manage them that's fine.
We can also create multiple renderes and objects of different type just tell that specific renderer how they are supposed to be rendered.
How do UI elements and "traditional" game objects differ from each other? The shader tells the gpu where and how it should be rendered and the class itself passes the data to the shaders and the gpu.
# No difference how the renderer renders the object
If we are using just a single `Render` function then a single renderer would make sense. However, iterating through all types of objects and calling `Render` on them may not the be the most performant solution with a single renderer.
Maybe we could create a `Renderer` class that handles rendering of objects of different types.
# Differences in the way objects need to be rendered
TBD: This is the part we need to investigate right now. If there are non, investigate the above and see what is the best / most performant / maintainable solution.
贡献指南
调研方向
未指定任何文件或测试。首先定位现有的 renderer、Render 方法、与 Renderable 相关的类型以及 shader 路径,然后比较一个 renderer 与分别用于 UI 和传统游戏对象的特定类型 renderer。完成的标准是记录一套基于性能、可维护性和开发者易用性而确定的架构。
由索引模型根据 Issue 内容生成。
评估
- 技术栈
- csharp
- 领域
- computer-graphics, game-dev
- Issue 类型
- 重构
- 难度
- 5/5
- 预计耗时
- 一周以上
- 活跃度
- 停滞
- 描述清晰度
- 需要澄清
- 新手友好度
- 25/100