AI 中文总结
该研究针对混合现实窗口排列的提案可视化，对比三种可视化技术与手动定位，开展用户研究后发现情境窗口预览布局更快，但参与者更偏好手动控制，为混合方法提供启示。
AI 中文摘要
自适应混合现实（MR）界面通常会代表用户优化窗口布局，却很少考虑个人偏好。一种有前景的替代方案是让用户参与其中，向用户展示布局提案供其选择，但这些提案应如何可视化却未得到充分探索。我们比较了三种窗口放置的提案可视化技术：情境图标预览（Situated Icon Preview）、情境窗口预览（Situated Window Preview）和3D预览，以手动定位（Manual Positioning）为基准。这些技术在细节程度和交互空间上下文的程度上存在差异。在一项被试内用户研究中，24名参与者在VR中完成了多阶段旅行规划任务，使用每种技术分别放置7个依次出现的窗口，我们重点关注预定义提案位置下的单窗口放置。我们测量了布局时间、布局变更次数、任务负荷、用户体验和偏好，并辅以半结构化访谈。尽管情境图标预览和情境窗口预览相比手动定位缩短了布局时间，且情境窗口预览比3D预览更快，但参与者更偏好直接手动控制。我们讨论了形成这种偏好的因素（感知控制、认知成本、熟悉度以及提案的信息量），并概述了混合方法的启示，即提案式建议与手动优化的组合是一种有前景的方案。
英文摘要
Adaptive mixed reality (MR) interfaces typically optimize window layouts on behalf of the user, with limited consideration for individual preferences. A promising alternative keeps users in the loop by presenting layout proposals for them to select from, but how these proposals should be visualized remains underexplored. We compare three proposal-visualization techniques for window placement, Situated Icon Preview, Situated Window Preview, and 3D Preview, against a Manual Positioning baseline. The techniques differ in level of detail and degree of interaction-space context. In a within-subjects user study, 24 participants completed a multi-stage trip-planning task in VR, individually placing seven sequentially introduced windows using each technique. We thereby focus on single-window placement under predefined proposal positions. We measured layouting time, number of layout changes, task load, user experience, and preference, complemented by semi-structured interviews. Although Situated Icon Preview and Situated Window Preview reduced layouting time compared to Manual Positioning - with Situated Window Preview also faster than 3D Preview - participants preferred direct manual control. We discuss the factors shaping this preference (perceived control, cognitive cost, familiarity, and informativeness of the proposal) and outline implications for hybrid approaches, as a promising combination of proposal-based suggestions with manual refinement.
