知识卡片
清理调试输出保留有用观测
无用途打印和中间展示可能淹没主线,可逐项判断后删除;必要质量反馈应迁到结构化结果、日志或测试。打印是可观察副作用,若消费方依赖输出就不能不核对契约直接称无行为变化。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 重构与技术债
原书出处 · 1 处
- · 查看原书依据
2. Remove print statements. This step removes noise and visual clutter and makes the next step—listing code smells—exponentially easier. If you see a print statement—or a variable left in the last line of the cell, which is displayed in the notebook output—consider whether it’s still serving a valid purpose. If it was added once upon a time to help with debugging and is now adding
参考来源
- 位置:《Effective Machine Learning Teams (David Tan, Ada Leung, David Colls)》第08章《重构与技术债》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Effective Machine Learning Teams (David Tan, Ada Leung, David Colls)/_epub-src/OEBPS/ch08.html)
- 结论依据:摘录中关于清理调试输出保留有用观测的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。
- 原始内容:2. Remove print statements. This step removes noise and visual clutter and makes the next step—listing code smells—exponentially easier. If you see a print statement—or a variable left in the last line of the cell, which is displayed in the notebook output—consider whether it’s still serving a valid purpose. If it was added once upon a time to help with debugging and is now adding