Guidelines for Communication, Development, Testing, and Post‑Release Practices

Analyzing many online incidents shows that most problems stem from a series of small, obvious mistakes rather than rare, complex bugs. The tendency to make these mistakes often comes from a lucky mindset or focusing only on short‑term gains.

Communication Principles

Use the appropriate communication tool for each scenario: instant messaging for simple, low‑cost messages; switch to voice or face‑to‑face when the issue is complex and text is insufficient. Separate important information into its own message to avoid being overlooked.

Avoid confirmation bias by actively seeking opposing viewpoints; they can reveal hidden risks or better solutions. Large‑scale recommendation algorithms can reinforce a single perspective, creating an information bubble.

Decision‑Making Guidance

Examples show that a single successful case should not be generalized. Decisions should be based on large‑scale statistics and context‑specific analysis rather than anecdotal evidence.

Requirements Review & Scheduling Principles

Everyone is a product manager: always question the necessity of a requirement and consider its real value. Technical solutions should be evaluated after the need is confirmed, not before. Align internal schedules before committing to external timelines to avoid mismatches.

Development Phase Principles

Maintain comprehensive documentation to reduce long‑term communication costs. Keep code readable, limit function length (<30 lines) and class size (<900 lines), and write comments before code. Use consistent, descriptive naming. Avoid clever tricks that sacrifice readability.

Commit messages act as an index; they should be concise, single‑purpose, and follow a fixed format (e.g., [branch][change]). Conduct code reviews as a fast feedback loop and treat reviewers’ comments as learning opportunities.

Testing Phase Principles

Cross‑team testing with diverse users uncovers hidden bugs. When facing difficult problems, consider whether the approach is wrong or knowledge is lacking; adopt a research mindset to fill gaps.

Post‑Release Principles

Online incidents require immediate mitigation, followed by root‑cause analysis and preventive measures. Treat each incident as a learning case, even if you are not directly responsible.

Overall, continuous accumulation of good habits—clear communication, disciplined documentation, readable code, thorough testing, and reflective post‑mortems—leads to higher efficiency and sustainable growth.