// the find
oldratlee/software-practice-thoughts
📚 🐣 软件实践文集。主题不限，思考讨论有趣有料就好，包含如 系统的模型分析/量化分析、开源漫游者指南、软件可靠性设计实践、平台产品的逻辑与执行… 🥤
A personal knowledge repo from a Chinese senior engineer collecting practice notes, slide decks, and Jupyter analysis notebooks on software design, reliability, caching, Git, and open-source strategy. It's written almost entirely in Chinese and targets developers who want opinionated engineering takes with some quantitative backing, not tutorials.
The load-balancing uniformity analysis (Poisson distribution modeling of task distribution skew) is genuinely useful thinking that most engineers skip. The CRC model for async request patterns (Completeness/Responsiveness/Correctness as trade-off axes) is a concrete mental framework, not hand-waving. The cache practice notes cover the actual failure modes — stampede, penetration, inconsistency — rather than just 'cache things for speed'. The compileflow code review is a real artifact showing how to evaluate an unfamiliar open-source codebase systematically.
Everything is in Chinese with no English translations, so its audience is immediately cut to a fraction of GitHub's user base. The repo is essentially a personal blog stored in git — no code, no runnable examples beyond the Jupyter notebooks, and the slide decks (.pptx) are opaque without actually opening them. The most recent substantive addition (concurrency orchestration, 2025-10) is a PDF/PPTX, not searchable text. For a repo categorized as 'best-practices', there's no way to know which practices are battle-tested vs. theoretical without reading Chinese fluently.