Back to Details

recognition-recents-and-suggestions

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Recents, frequents, and contextual suggestions

最近使用项、常用项与上下文建议

Recognition is fastest when the user doesn't even have to scan a long list — when the system anticipates and surfaces the likely options first. Recents, frequents, and predictive suggestions all leverage this: the user's likely target is at the top, often without typing anything.
当用户甚至无需扫描长列表时，识别速度最快——此时系统会预判并优先展示可能的选项。最近使用项、常用项和预测性建议都利用了这一点：用户的目标选项通常位于顶部，往往无需输入任何内容。

Patterns

设计模式

Recents

最近使用项

A list of items the user has recently interacted with. Surfaced at the top of pickers, navigation, or empty search inputs.
html
<combobox label="Recipient">
  <input placeholder="Search recipients..." />
  <listbox>
    <group label="Recent">
      <option>Maria Mendoza (last sent: yesterday)</option>
      <option>Marketing Team (last sent: 3 days ago)</option>
    </group>
    <group label="All">...</group>
  </listbox>
</combobox>
Most users compose for a small set of recipients repeatedly; recents collapse the recall task.
用户最近交互过的项目列表。展示在选择器、导航栏或空搜索框的顶部。
html
<combobox label="Recipient">
  <input placeholder="Search recipients..." />
  <listbox>
    <group label="Recent">
      <option>Maria Mendoza (last sent: yesterday)</option>
      <option>Marketing Team (last sent: 3 days ago)</option>
    </group>
    <group label="All">...</group>
  </listbox>
</combobox>
大多数用户会反复给一小部分收件人发送内容；最近使用项简化了回忆任务。

Frequents

常用项

Items used most often (regardless of recency). Useful when usage is clustered around a small set but not necessarily recent.
Frequently used apps:
  • Email
  • Slack
  • Code editor
  • Browser
A common laptop dock pattern.
使用频率最高的项目（无论时间远近）。适用于使用集中在一小部分项目但不一定是最近使用的场景。
Frequently used apps:
  • Email
  • Slack
  • Code editor
  • Browser
这是笔记本电脑 dock 的常见设计模式。

Recommended / suggested

推荐/建议项

System-predicted likely options based on context. Examples:
  • A "for you" feed.
  • "People you may know."
  • "Suggested replies" in messaging.
  • "Suggested tags" when categorizing.
Recommendations work when the prediction is good. Bad recommendations (irrelevant, wrong) are worse than none — they distract and erode trust.
系统根据上下文预判的可能选项。示例包括：
  • “为你推荐”信息流
  • “你可能认识的人”
  • 消息中的“建议回复”
  • 分类时的“建议标签”
当预测准确时，推荐功能才有效。糟糕的推荐（无关、错误）比没有更糟——它们会分散注意力并削弱用户信任。

Pinned / favorites

固定/收藏项

User-curated frequently-accessed items. Less algorithmic than recents/suggestions; user-explicit.
Pinned:
  ★ Q4 Planning Doc
  ★ Team OKRs
  ★ Customer feedback dashboard
Combine with recents and suggestions for a complete fast-access surface.
用户自行整理的常用访问项目。相比最近使用项/建议项，算法参与度更低，是用户明确指定的内容。
Pinned:
  ★ Q4 Planning Doc
  ★ Team OKRs
  ★ Customer feedback dashboard
将固定项与最近使用项、建议项结合，可打造完整的快速访问界面。

Smart defaults

智能默认值

Pre-fill fields with predicted values based on context (signed-in user, recent inputs, time of day, location).
html
<form>
  <label>Country
    <select name="country">
      <option value="US" selected>United States</option>
      <!-- selected because of user's IP location -->
    </select>
  </label>
</form>
The user can change but rarely needs to.
根据上下文（登录用户、最近输入、时间、位置）预填充字段的预测值。
html
<form>
  <label>Country
    <select name="country">
      <option value="US" selected>United States</option>
      <!-- selected because of user's IP location -->
    </select>
  </label>
</form>
用户可以更改，但很少需要这么做。

When recents/suggestions hurt

最近使用项/建议项的弊端

  • When the prediction is bad. Wrong recents distract; the user has to filter past them.
  • When privacy matters. Recents reveal user history; in shared-device contexts this can leak information.
  • When the option set is critical to the task. A "recent" suggestion in a destructive action might bias the user toward the wrong choice.
For high-stakes actions, present the full set without privileging recents.
  • 预测不准确时：错误的最近使用项会分散注意力，用户必须筛选掉它们。
  • 涉及隐私时：最近使用项会暴露用户历史；在共享设备场景下可能泄露信息。
  • 选项集对任务至关重要时：破坏性操作中的“最近”建议可能会引导用户做出错误选择。
对于高风险操作，请展示完整选项集，不要优先显示最近使用项。

Privacy and recents

隐私与最近使用项

Recents reveal user activity to anyone with screen access. Considerations:
  • Don't surface recents on shared/public devices unless explicitly opted in.
  • Provide a "clear recents" option.
  • Don't leak across tenants (a recent in workspace A shouldn't appear when the user switches to workspace B).
  • Be cautious with sensitive contexts (health apps, finance apps, dating apps).
最近使用项会向任何能访问屏幕的人暴露用户活动。需考虑以下几点：
  • 不要在共享/公共设备上展示最近使用项，除非用户明确选择开启。
  • 提供“清除最近使用项”的选项。
  • 不要跨租户泄露（工作区A的最近使用项不应在用户切换到工作区B时显示）。
  • 谨慎处理敏感场景（健康应用、金融应用、约会应用）。

Anti-patterns

反模式

  • Stale recents that include items the user no longer cares about, never expiring.
  • Recents that span privacy boundaries (work email recents in personal context).
  • Suggestions that don't update as the user's behavior changes.
  • Surfacing recents in wrong contexts (showing "recent files" on a different user's account).
  • 过时的最近使用项：包含用户不再关心的项目且永不过期。
  • 跨越隐私边界的最近使用项（工作邮箱的最近使用项出现在个人场景中）。
  • 不随用户行为变化更新的建议项。
  • 在错误上下文展示最近使用项（在其他用户账户上显示“最近文件”）。

Heuristics

启发法

  1. The "would the user pick this without typing?" check. For each picker, ask: in the median case, can the user get to their target without typing? If yes, recents are doing their job.
  2. The recents-quality audit. Sample your recents lists. Are they actually relevant to current intent?
  3. The privacy review. What do recents reveal? Should they be hidden by default in some contexts?
  1. “用户无需输入就能选中吗？”检查：对于每个选择器，问自己：在一般情况下，用户无需输入就能找到目标吗？如果是，说明最近使用项发挥了作用。
  2. 最近使用项质量审核：抽样检查最近使用项列表，它们是否与当前意图真正相关？
  3. 隐私审查：最近使用项会暴露什么信息？在某些场景下是否应默认隐藏？

Related sub-skills

相关子skill

  • recognition-over-recall
     (parent).
  • recognition-pickers-and-palettes
     — picker patterns recents augment.
  • satisficing
     — recents enable satisficing by surfacing acceptable options first.
  • hicks-law-defaults
     — defaults and recents both reduce decision cost.
  • recognition-over-recall
    （父范畴）。
  • recognition-pickers-and-palettes
    ——最近使用项所增强的选择器模式。
  • satisficing
    ——最近使用项通过优先展示可接受的选项，实现“满意即可”的决策方式。
  • hicks-law-defaults
    ——默认值和最近使用项都能降低决策成本。