Stuff I’ve Seen: A System for Personal Information Retrieval and Re-Use Susan Dumais, Edward Cutrell, JJ Cadiz, Gavin Jancke, Raman Sarin, Daniel C. Robbins Microsoft Research One Microsoft Way Redmond, WA 98052 USA [sdumais; cutrell; jjcadiz; gavin; ramans; dcr]@microsoft.com ABSTRACT Most information retrieval technologies are designed to facilitate information discovery. However, much knowledge work involves finding and re-using previously seen information. We describe the design and evaluation of a system, called Stuff I’ve Seen (SIS), that facilitates information re-use. This is accomplished in two ways. First, the system provides a unified index of information that a person has seen, whether it was seen as email, web page, document, appointment, etc. Second, because the information has been seen before, rich contextual cues can be used in the search interface. The system has been used internally by more than 230 employees. We report on both qualitative and quantitative aspects of system use. Initial findings show that time and people are important retrieval cues. Users find information more easily using SIS, and use other search tools less frequently after installation. General Terms Algorithms, experimentation, human factors. Keywords Personal information management, user interfaces, user studies, interactive information retrieval. INTRODUCTION Most information retrieval tools, like popular web and intranet search engines, are designed to facilitate information discovery. Given a short query, they do a remarkable job of finding relevant materials using a variety of content, anchor text, link and popularity cues. However, much knowledge work involves integrating and re-using information that has previously been created or accessed. For example, writing a presentation or paper may involve some web searching, but it also involves pulling together information from existing information sources like documents, spreadsheets, data analyses, email messages, etc. Studies have shown that 58-81% of web pages accessed were re-visits to pages previously seen [9,23,29]. Similar re-access patterns have been observed in usage of Unix commands [15], library book borrowing [7], and human memory [3]. We developed a system called Stuff I’ve Seen (SIS) that makes it easy for people to find information they have seen before. Two key aspects of the SIS design support this. First, the system provides a unified index of information that a person has seen on their computer, whether the information was an email, web page, document, media file, calendar appointment, etc. Today, people have to manage several different organizations of information – e.g., the file system hierarchy for files, the email folder hierarchy for email, favorites or history for web pages. With SIS, all of these sources are integrated into a single index regardless of what form the information originated. Second, because a person has seen the information before, rich contextual cues such as time, author, thumbnails and previews can be used to search for and present information. In contrast, web search results lack personal context, so rank is about the only reasonable alternative for ordering results. RELATED WORK Vannevar Bush’s vision of memex [8], “a device in which an individual stores all his books, records, and communications, and which is mechanized so that it may be consulted with exceeding speed and flexibility” captures the essence of SIS. The details of implementation are obviously very different, but the functionality is much the same as what we have developed. Jones and Thomas [18] surveyed people’s use of new personal information management technologies and found low adoption rates. They speculated that the limited applicability of new technologies, each focusing on a limited conception of personal information, was the main reason for slow adoption. Erickson [12] developed a new personal notebook application. He described field observations of usage patterns as well as the co-evolution of the system and work practices. Our focus was not to create a new information management application but rather to develop a unified search interface to existing content sources. Several groups have looked at methods for improving access to subsets of information. Malone [21] observed Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. SIGIR’03, July 28 – August 1, 2003, Toronto, Canada. Copyright 2003 ACM 1-58113-646-3/03/0007…$5.00.how people organize paper materials and suggested how computer systems could better support these activities. Barreau and Nardi [4] examined how people manage files on their computer. Whittaker and Sidner [30] described the email overload problem and linked it to difficulties that people have with deferred processing and classification of items. Jones et al. [17] conducted detailed observations of the methods that people use to organize web pages for re-use and developed a functional analysis to show how the techniques people use depend on anticipated re-access needs. Several groups developed systems to improve re-access to web pages, including the use of rich graphical representations [10], integration of back, history and favorites [19], predictive models of information needs [24], and full text indexing of pages [22]. These investigations all focused on improving access to information within a single content type. Some research systems provide access to more than a single type of personal information. Haystack [1,16] is a personal store that supports annotations and collections. The initial version worked with input from the browser and editor, although extensions to email and files were planned. MyLifeBits [14] is similar to SIS but focuses on multimedia files and support for rich annotations. Lifestreams was developed as a replacement for the desktop metaphor [13]. It provides a single time-ordered stream of electronic information, and supports searching, filtering and summarization. Several commercial products have functionality to index some personal information. Microsoft’s Indexing Services and Apple’s