One of the most common ways in which results are displayed by an information retrieval system is in the form of a list, in which the most relevant results appear in the first positions. Today's large screens, however, allow one to create more complex displays of results, especially in cases such as image retrieval, in which each unit returned is fairly compact. For these layouts the simple list model is no longer valid, since the relations between the slots in which the results are placed do not form a sequence, that is, the relation among them is no longer that of a total order. In this paper we model these layouts as partial orders and show that a “stalwart display” property (a layout in which items’ relevance is unambiguously conveyed by their display position) can be obtained only in the case of lists. For the other layouts, we define two classes of representation functions: “safe” functions (which display results without adding spurious structure) and “rich” functions (which do not drop any structure from the result set), as well as an algorithm to optimally display fully ordered result sets in arbitrary display layouts.

中文翻译：

---

在非列表显示布局中将评级项目分配给位置

信息检索系统显示结果的最常见方式之一是列表形式，其中最相关的结果出现在第一位置。但是，当今的大屏幕允许人们创建更复杂的结果显示，尤其是在诸如图像检索的情况下，其中返回的每个单元都非常紧凑。对于这些布局，简单列表模型不再有效，因为放置结果的插槽之间的关系不会形成序列，也就是说，它们之间的关系不再是总顺序的关系。在本文中，我们将这些布局建模为部分订单，并表明只有在列表的情况下，才能获得“稳健的显示”属性（一种布局，其中项目的相关性通过其显示位置明确传达）。对于其他布局，