如何理解Reporting a suspected platform issue？

返回 技术表达与英语 目录

标准回答

I would provide the exact device and software versions, reproduction steps, the affected time range, and traces from both a normal and a slow run. I would clearly separate observations from hypotheses.

For example, I might say: The client waits for 160 milliseconds, while the server has only 9 milliseconds of CPU execution. We still need to identify the queueing and blocking time. Could we inspect the server thread and its downstream dependencies together?

中文要点：160ms和9ms来自本 Wiki练习，实际交流必须换成真实数据。用observations与hypotheses区分证据和推断；提出具体合作任务，而非简单说platform is slow。

追问与回答

What if the other team disagrees

I would propose a small experiment that can distinguish the two explanations, and agree on the data needed to confirm or reject each hypothesis.

内容来源

• Android 与 AI 工程 Wiki · 原条目 q102