The ReSearch Engine – Helping People Return to Information on the Web.pdfVIP

The Re:Search Engine – Helping People Return to Information on the Web Jaime Teevan Massachusetts Institute of Technology, CSAIL Cambridge, MA 02 139 USA teevan@ ABSTRACT that even among this technically savvy population, partici- Re-finding information is commonly cited as a problem on pants preferred navigating to their search target along the Web. One reason re-finding is hard is that while people known paths over j umping directly to it [10]. This pre- rely on context to return to information (e.g., by following ferred search strategy is fragile, failing if any part along the the original path taken to it), the Web makes no guarantee known path changes. A naturalistic study analyzing in- that the context will remain static. The Re:Search Engine is stances mined from the Web where people expressed diffi- designed to help people return to information in dynamic culty re-finding found that lost information was commonly environments like the Web by maintaining consistency in described using the path originally taken to find it [12]. the results it returns. For example, if Connie, while looking …BUT THE WEB CHANGES to purchase a Global Positioning System, found several she Despite the importance of consistency in re-finding, infor- liked via a search for “GPS”, she would expect to be able to mation on the Web frequently changes. For example, use the same query to locate the same system