Back to Details

review-ui

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Goal

目标

Assess an interface and return evidence-backed findings without changing it.
评估界面并在不修改界面的情况下返回有证据支持的问题分析。

Definition of Done

完成标准

  • Scope, candidate, environment, task, and evidence types are identified.
  • The main journey and platform-relevant states are assessed or explicitly excluded.
  • Findings distinguish observations, preferences, and hypotheses and include priority, location, impact, correction, and acceptance check.
  • Unobserved interaction and accessibility gaps are explicit.
  • Current validation accepts the report's declared coverage; no required finding remains.
  • 明确范围、候选对象、环境、任务和证据类型。
  • 评估主要用户流程和与平台相关的状态，或明确将其排除在评估范围之外。
  • 区分观察结果、偏好和假设，并在问题分析中包含优先级、位置、影响、修正建议和验收检查。
  • 明确指出未观察到的交互和无障碍缺口。
  • 当前验证已接受报告声明的覆盖范围，且没有遗留任何必须处理的问题。

Boundaries

边界

  • Do not repair the interface, install anything, create hosted issues, or change external state. A missing observation is incomplete, not a pass; N/A needs a reason.
  • 不得修复界面、安装任何内容、创建托管问题或更改外部状态。缺少观察结果应视为覆盖不完整，而不是通过；标记为 N/A 时必须说明原因。

Workflow

工作流

  1. Resolve platform, screens, audience, task, candidate/environment, rules, permitted read-only evidence, bounded scope, and validation criteria.
  2. Trace the main journey and applicable layout, content, feedback, loading, empty, error, validation, disabled, and success states.
  3. Check platform-relevant interaction and accessibility. For web, inspect semantics, names, keyboard/focus, contrast, and motion; when an expanded assessment is required, use UI accessibility in review mode with the scope, candidate, and evidence. Use Web performance only when requested, in review mode. Distinguish source, screenshot, automated, keyboard, and assistive evidence.
  4. Return strengths and ordered findings with evidence, priority, location, state or viewport, impact, correction, and acceptance check.
  5. For a substantive report, run Quality validation in read-only mode with the scope, candidate, and evidence. Revise only the report or gather permitted evidence, then renew affected verdicts until the declared coverage is accepted.
  1. 确定平台、界面、受众、任务、候选对象/环境、规则、允许使用的只读证据、有限范围和验证标准。
  2. 梳理主要用户流程，以及适用的布局、内容、反馈、加载、空状态、错误、校验、禁用和成功状态。
  3. 检查与平台相关的交互和无障碍表现。对于 Web 界面，检查语义、名称、键盘操作/焦点、对比度和动效；需要扩展评估时，使用 UI accessibility 并以审查模式传入范围、候选对象和证据。仅在明确要求时，以审查模式使用 Web performance。区分源码、截图、自动化、键盘和辅助技术证据。
  4. 返回优点和按优先级排序的问题分析，其中包含证据、优先级、位置、状态或视口、影响、修正建议和验收检查。
  5. 对于实质性报告，使用 Quality validation 并以只读模式传入范围、候选对象和证据。只能修改报告或收集允许的证据，然后重新验证受影响的结论，直至报告声明的覆盖范围通过验证。