A failed test is a clue. Follow the evidence.
A practical way to move from a red CI job to a defensible root cause: separate observations, hypotheses, and proof.
FIELD NOTES FROM THE WORKBENCH
Engineering,
with curiosity.
Storage systems. Reliable tests. AI that holds up.
A notebook for making complex engineering
a little clearer.
SystemsIntelligenceCode
A GOOD PLACE TO STARTFEATURED NOTE
A failed test is a clue.
Follow the evidence.
A practical way to move from a red CI job to a defensible root cause: separate observations, hypotheses, and proof.
Read the noteTHE NOTEBOOK
All notes.
Practical notes on engineering and English learning.
approach 精读：处理方式、着手处理与接近
从一个单词到能用的表达：分清名词 approach to doing 与动词 approach a problem，再用例句、易错对比和自测巩固。
迭代与递归学习系列：42 个知识点的详细讲解与复盘索引
五篇主线，从循环状态到调用栈、尾递归、递归树与显式栈。每个知识点都有详细章节，区分原文、推导和补充，方便逐项复盘。
理解迭代：for、while 与嵌套循环的复杂度
从迭代器和条件检查理解循环机制，用操作计数澄清“for 必须预知次数”和“多嵌套一层就升一阶”这两种常见误解。
理解递归：从函数调用到逐层返回
通过阶乘、调用栈交互与实际日志，理解每次调用独立的局部状态、返回后如何继续，以及多层递归调用如何共存。
理解尾递归：累加器、执行过程与 Python 的优化边界
从待完成的加法判断尾递归，用状态表和不变量解释累加器，再区分代码形式、尾调用优化与 CPython 的实际行为。
理解递归树：斐波那契、调用次数与记忆化
一棵树记录全部调用，一条路径决定峰值深度。用斐波那契的九节点图和实际计数，看清重复子问题，再比较缓存与循环。
递归如何变成迭代：显式栈、继续执行的位置与算法取舍
从求和到树的后序遍历，观察显式栈怎样保存参数和待完成动作；比较状态数量、输出顺序、深度约束与可读性。
Test the recovery, not just the interruption.
Fault injection gets interesting after the fault. Define what recovery means from the host, data, and system perspectives.
One lucky run is not an agent evaluation.
Turn an impressive demo into a repeatable evaluation with explicit outcomes, multiple trials, and readable traces.
Treat a prompt change like a release.
Version the behavior you expect, compare it against a baseline, and keep enough evidence to understand regressions.
Python 的 while 与 for：三个循环示例
从动态步长、条件终止到用户交互，理解 while 与 for 的控制机制，以及两者的适用场景。
A lock protects an invariant, not a variable.
Think about the state that must stay consistent, then make the whole read–check–write operation participate in the same protocol.
⌕
No notes found.
Try another phrase or explore all topics.
— You’re all caught up. There’s always more to understand. —
THE COMMON THREAD
Good engineering starts
with a better question.
Observe carefully.
Make the idea testable.
Let the evidence lead.