150 comments • Comments Link

• 一些开发者在系统编程上更偏好 Odin 而非 Rust 或 Zig，原因是 Odin 与 C 的互操作性更好，不像 RAII 那样带来复杂的内存管理负担，同时认知开销更低。

• RAII 的批评者认为，将内存分配与对象初始化耦合会造成"death by a thousand cuts"，过多的临时分配会损害性能，因此他们更倾向于手动内存管理或游戏开发中常见的基于 arena 的内存池。

• 支持 RAII 和现代 C++ 的人则认为，只要这些模式与 vectors 等容器正确搭配使用，它们是高效的；针对 RAII 的"性能"指责往往更多是为选择其它方案寻找借口，而非基于数据的必然结论。

• 围绕 Odin 、 Zig 、 Jai 等语言的讨论，源自人们希望找到一种能替代 C++ 的语言：既保有低级别控制，又不带着 C++ 历史遗留的复杂性和"丑陋"特性。

• 有人认为 Rust 受制于一种较为僵化和完美主义的文化，这延缓了关键系统特性的稳定（例如可插拔的 allocator），令那些需要精细内存控制的工程师沮丧。

• 尽管缺乏像成熟语言那样丰富的生态与工具支持，Odin 在 STM32 microcontrollers 等资源受限环境中已展示出实际可用性。

• 新语言的传播往往靠"怪癖"吸引注意，而不是通过明确的使用场景来说明定位，这会让潜在采用者不清楚它是面向 data-oriented 、偏向图形，还是通用的系统工具。

• LLMs 的兴起给语言采纳带来了新的障碍：开发者越来越倾向于优先学习那些 LLMs 能高质量生成代码的语言，这会降低投入时间学习小众或新语言的回报率。

• 为 LLMs 专门设计的语言可能需要强调高表达能力、减少 token 数量并保持逻辑一致性，以尽量降低顺序化代码生成时误差的累积效应。

• 一门语言是否成功，往往更多取决于用它构建出的真实且高质量的项目，这些项目能清晰地展示出权衡点与实际优势，而不是单靠某些特性本身。

这场讨论反映出开发者群体内在的深刻分歧：一方面优先强调对硬件资源的显式、手动控制，另一方面则偏好 RAII 、 borrow checker 之类的高级抽象。尽管大家对 C++ 的历史复杂性和 Rust 演进速度迟缓普遍感到疲惫，但对于像 Odin 这样的"data-oriented"新语言究竟是代表范式转移，还是仅仅在游戏开发等特定领域里成为专业工具，目前仍无定论。总体上，这场争论凸显了语言设计者的审美与哲学取向，与专业工程师在 AI 辅助开发时代对由生态驱动的实际需求之间的张力——在这个时代，学习新语法的价值主张本身也在发生变化。 • Odin is preferred by some developers over Rust and Zig for systems programming due to its superior C interoperability, lack of complex memory management baggage like RAII, and lower cognitive overhead.

• Critics of RAII argue that coupling memory allocation with object initialization leads to "death by a thousand cuts," where excessive temporary allocations degrade performance, favoring instead manual memory management or arena-based pools common in game development.

• Proponents of RAII and modern C++ argue that these patterns, when used correctly with containers like vectors, are performant, and that the "performance" critique of RAII is often a rationalization rather than a data-backed technical necessity.

• The debate surrounding languages like Odin, Zig, and Jai is driven by a desire for an alternative to C++ that maintains low-level control without the historical complexity and "ugliness" of C++ features.

• Rust is viewed by some as suffering from a rigid, perfectionist culture that slows the stabilization of critical systems features, such as custom/pluggable allocators, leading to frustration among engineers who need fine-tuned memory control.

• Odin demonstrates practical utility in resource-constrained environments like STM32 microcontrollers, though it lacks the extensive ecosystem and tooling support found in more established languages.

• The marketing of new programming languages often relies on "quirks" rather than showcasing specific use cases, which can confuse potential adopters about a language's true purpose, whether it is data-oriented, graphics-focused, or a general-purpose systems tool.

• The rise of LLMs introduces a new hurdle for language adoption, as developers increasingly prioritize languages that LLMs can generate code for effectively, potentially reducing the ROI for learning niche or newer languages.

• Designing a language specifically for LLMs might involve focusing on high expressivity, reduced token count, and logical consistency to minimize the error-compounding effects of sequential code generation.

• A language's success depends less on its specific features and more on the tangible, high-quality projects built with it that demonstrate clear trade-offs and real-world advantages.

The discussion reflects a deep divide between developers who prioritize explicit, manual control over hardware resources and those who favor high-level abstractions like RAII and borrow checkers. While there is a palpable fatigue with the complexity of C++ and the perceived slow evolution of Rust, there is no consensus on whether newer "data-oriented" languages like Odin represent a paradigm shift or simply a specialized tool for specific domains like game development. Ultimately, the conversation highlights a tension between the aesthetic and philosophical preferences of language designers and the practical, ecosystem-driven needs of professional software engineers in an era where AI-assisted development is shifting the value proposition of learning new syntax.