This paper discusses a navigation behavior on Internet information services, in particular the World Wide Web, which is characterized by pointing out of information using various communication tools. We call this behavior social navigation as it is based on communication and interaction with other users, be that through email, or any other means of communication. Social navigation phenomena are quite common although most current tools (like Web browsers or email clients) offer very little support for it. We describe why social navigation is useful and how it can be better supported in future systems. We further describe two prototype systems that, although originally not designed explicitly as tools for social navigation, provide features that are typical for social navigation systems. One of these systems, the Juggler system, is a combination of a textual virtual environment and a Web client. The other system is a prototype of a Webhotlist organizer, called Vortex. We use both systems...

This dissertation examines the use of adaptive methods to automatically improve the performance of ranked text retrieval systems. The goal of a ranked retrieval system is to manage a large collection of text documents and to order documents for a user based on the estimated relevance of the documents to the user's information need (or query). The ordering enables the user to quickly find documents of interest. Ranked retrieval is a difficult problem because of the ambiguity of natural language, the large size of the collections, and because of the varying needs of users and varying collection characteristics. We propose and empirically validate general adaptive methods which improve the ability of a large class of retrieval systems to rank documents effectively. Our main adaptive method is to numerically optimize free parameters in a retrieval system by minimizing a non-metric criterion function. The criterion measures how well the system is ranking documents relative to a target ordering, defined by a set of training queries which include the users' desired document orderings. Thus, the system learns parameter settings which better enable it to rank relevant documents before irrelevant. The non-metric approach is interesting because it is a general adaptive method, an alternative to supervised methods for training neural networks in domains in which rank order or prioritization is important. A second adaptive method is also examined, which is applicable to a restricted class of retrieval systems but which permits an analytic solution. The adaptive methods are applied to a number of problems in text retrieval to validate their utility and practical efficiency. The applications include: A dimensionality reduction of vector-based document representations to a vector spa...

Prototypes are widely recognized to be a core means of exploring and expressing designs for interactive computer artifacts. It is common practice to build prototypes in order to represent different states of

Systems for fast search of personal information are rapidly becoming ubiquitous. Such systems promise to dramatically improve personal information management, yet most are modeled on Web search in which users know very little about the content that they are searching. We describe the design and deployment of a system called Phlat that optimizes search for personal information with an intuitive interface that merges search and browsing through a variety of associative and contextual cues. In addition, Phlat supports a unified tagging (labeling) scheme for organizing personal content across storage systems (files, email, etc.). The system has been deployed to hundreds of employees within our organization. We report on both quantitative and qualitative aspects of system use. Phlat is available as a free download at

This study deals with the problems users encounter in their daily work with computers and the typical practices that they employ. Sixteen daily computer users were interviewed about their habits and problems that they encountered during document classification and retrieval. For both these areas, we provide an overview of identified user practices and a citation-based analysis of the problems users encountered, including those related to the use of the screen real estate (the actual desktop). Two types of problems were identified: (1) Problems that concern the actual use of the system installed on the computer. (2) Problems that arise when people realise that they are using a system that does not allow for the desired work or organizational functions sought. We were able to show that skill continues to be an important factor with respect to the ease of using today’s systems. We suggest the following necessary improvements for the evolution of personal information systems: A storage facility that represents the user’s view of information; replacing pure technical file metadata with more user-friendly attributes; and introduction of annotations as a new information type.

This paper describes V-Twin, an information access toolkit designed to provide indexing and search capabilities for a variety of applications. We discuss the phenomenon of very short queries generated by users of interactive search services, and summarize a new technique we are using in V-Twin to handle these queries more effectively. We then present some results based on V-Twin's performance at the TREC-5 ad hoc task. VTwin achieved a high level of performance despite having much lower index overhead and memory footprint than other systems participating in TREC. 1. V-Twin Background V-Twin is the code name of the Apple Information Access Toolkit (AIAT), originally created in Apple Research Labs and now available for licensing by Apple developers. The toolkit consists of a library of C++ classes designed to be linked to a front-end application. V-Twin is based on the standard vector model of information retrieval, with a variant of tf x idf weighting and length normalization. V-Twin ...

{trigg, blomberg, suchman}®'pare.xerox.com •• Abstract. This paper reports on a work-oriented design project concerned with the question of how to migrate shared, workgroup document collections currently kept on paper online. Based in a civil engineering work group, the focus of our project is a document collection called the "project files, " a heterogeneous mix of documents that serve as an ongoing resource for the group during a project's course as well as an archival record at its completion. We describe the dynamics of the standardized classification scheme in use for the project files, existing practices of document filing including routine troubles, and the prototype developed to move the project files online. The latter includes a configuration of hardware and software along with associated practices of document scanning, coding and search. We conclude with some reflections on the difficulties of maintaining alignment across paper and digital media in the migration to online document collections, and with a summary of the questions posed and answers provided by our project.

Satisfying non-trivial information needs involves collecting information from multiple resources, and synthesizing an answer that organizes that information. Traditional recall/precision-oriented information retrieval focuses on just one phase of that process: how to efficiently and effectively identify documents likely to be relevant to a specific, focused query. The TREC Interactive Track has as its goal the location of documents that pertain to different instances of a query topic, with no reward for duplicated coverage of topic instances. This task is similar to the task of organizing answer components into a complete answer. Clustering and classification are two mechanisms for organizing documents into groups. In this paper, we present an ongoing series of experiments that test the feasibility and effectiveness of using clustering and classification as an aid to instance retrieval and, ultimately, answer construction. Our results show that users prefer such structured p...

Scrollbars provide a simple way to traverse an information space, but they provide little data about the actual contents of the space. Of the many visualization techniques that have been proposed, few have maintained the simple functionality of the scrollbar while showing improved performance on typical scrollbar tasks. This paper presents two enhancements to the scrollbar, a mural bar and a pile bar, which encode data about the information space contents into the trough of the bar. Results from an experiment suggest that these new devices lead to improved user performance on several common scrollbar tasks.

The typical computer application follows a set pattern: it is executed by a user, it provides responses to commands issued by the user, and it is terminated by the user when it is no longer needed. However, the advent of networked computers and the World-Wide Web has changed the way many applications work. Agents and similar programs constantly access online information databases and identify items that may be of interest to their users. The problem lies in communicating the collected information. Most existing widgets cannot take full advantage of the constant information flux generated by these programs. They work best as part of a full-screen display, they do not react well to changes in information, or they cannot attract the attention of users when new information arrives. This paper explores these problems and describes anumber of interactors that may help alleviate them.