Document Triage in the Information Seeking Process

当用户搜索信息时，他们会快速阅读文档以评估其有用性。这种简短的行动/称为文件分类/通常会进行多次，以寻找足够的信息来满足一个特定的需求。用户可以尝试通过诸如Google之类的系统进行不同的搜索，或者浏览文档目录，但是在任何一种情况下，他们都需要定期打开诸如网页或电子书（PDF）之类的数字文档，以评估其真实的相关性。实验证据表明，在这些简短的审查文件中作出的判断发挥了重要作用，在用户的信息寻求的整体成功。在拒绝相关材料和选择不相关的材料进行进一步的阅读时，经常会出现错误。对于大多数搜索而言，只有用户在浪费对无用材料的仔细阅读和对定位相关信息的重复尝试中的大量进一步努力才最终实现足够成功的结果。这显然是低效的。本项目将仔细研究用户和文档显示软件之间的交互对用户在此过程中的决策的影响。在优化人类努力成果的目标驱动下，它将研究在文档的初始阅读过程中影响相关性决策质量的积极和消极因素。虽然用户阅读文档主要是为了评估其相关性，但同时也会触发许多其他事件。可以识别新的信息需求，可以意外地找到另一个问题的答案，并发现新搜索的替代术语。这些次要活动可能对用户的搜索计划产生巨大影响，但对用户相关需求的了解有限。在物理环境中，这些目标通常是使用笔记本、纸片和人类记忆来跟踪的。数字信息搜索工具无法替代这些工具，因此现有的实践通常与电子搜索一起延续。这些改进的知识将用于设计和提供有用的数字化工具，以支持这项工作，这些工具将与搜索引擎和文件组织软件等现有的信息搜索设施相结合。可据以建立良好的人类行为模式的相关数据十分短缺。虽然相关的领域，如详细的阅读较长的文本，已被研究，这些发现，以及数据支持，不能直接应用。用户在物理环境中对人类搜索者的观察提供了在使用计算机软件时可能被证明是错误的见解。纸质图书的使用者本身也受到他们所使用的图书属性的限制。新的方法可能是数字化的，而纸张无法提供，由于两种媒介之间的差异，基于纸张的行为的数字复制可能证明是无效的。目前，软件设计师只能凭直觉为文档阅读器应用程序使用最好的交互。这些软件工具最初的设计往往是为了支持文件的下载和打印，而且已经包括了支持系统阅读较长文件的功能。科学家目前缺乏理论上的见解，建议有效的解决方案，以支持分类阅读的文件。文件分类在决策过程中发挥着核心作用，推动信息寻求。更深入地了解文件分类将作出重大贡献的不断发展的科学理解的信息寻求过程作为一个整体

When users are searching for information, they quickly read documents to assess their usefulness. This brief action / termed document triage / will often be undertaken many times in the pursuit of finding enough information to satisfy one particular need. The user may attempt different searches through a system such as Google, or browse a catalogue of documents, but in either case, they will regularly need to open a digital document such as a web page or electronic book (PDF) to evaluate its real relevance. Experimental evidence demonstrates that the judgements made in these short reviews of documents play a significant role in the overall success of a user's information seeking. There are frequent errors made in rejecting relevant material, and selecting irrelevant material for further reading. Only substantial further effort by the user, in wasted close reading of unhelpful material and repeated attempts to locate relevant information will finally achieve a sufficiently successful outcome for most searches. This is clearly inefficient. This project will scrutinise the impact that the interaction between the user and the document display software has on the user's decision making during this process. Driven by the goal to optimise the outcome of human effort, it will examine the positive and negative factors that affect the quality of relevance decisions during the initial reading of a document.While users are reading a document primarily to assess its relevance, there are many other events that can be triggered at the same time. New information needs can be identified, answers to another question can be unexpectedly found, and alternative terminology for a new search discovered. These secondary activities can have dramatic impacts on the user's search plans, but there is limited understanding of the related needs of users. In physical environments, these goals are usually tracked using notebooks, scraps of paper and human memory. Digital information-seeking tools provide no replacement for these tools, and thus existing practice is usually continued alongside electronic searching.Motivated by this gap, this research will refine our understanding of goal-tracking during information seeking. This improved knowledge will be used to design provide useful digital tools to support this work that integrate with existing information seeking facilities such as search engines and document organisation software.There are a number of challenges to address in achieving these objectives. Relevant data, from which good models of human behaviour can be built, is in short supply. Though related areas, such as the detailed reading of longer texts, have been studied, these findings, and the data that underpins them, cannot be applied directly. User observations of human searchers in physical environments provide insights that may prove false when using computer software. The users of paper-based books are themselves limited by the properties of the books they use. New methods may be available digitally that paper cannot supply, and digital reproductions of paper-based behaviour may prove ineffective due to differences between the two mediums.At present, software designers can only intuit the best interactions for the document reader applications to use. These software tools were often originally designed to support the download and printing of documents, and already include features that support the systematic reading of longer documents. Scientists currently lack the theoretical insights to recommend effective solutions for supporting the triage reading of documents.Document triage plays a central role in the decision-making processes that drive information seeking. A deeper understanding of document triage will make a significant contribution to the ongoing development of scientific understanding of the information seeking process as a whole