The web is large and complex, and in the process of navigating it, we often lose our way. Research trailing is a method to organize web contents that we have spent some effort on into distinct research sessions. Research trails are automatically constructed by filtering and organizing users’ activity history, using a combination of semantic and temporal criteria for grouping similar web activity. The design of research trails was informed by an ethnographic study of ordinary people doing research on the web; it addresses the specific challenges of establishing and maintaining context when the research process is fragmented and the research question is still in formation. This paper motivates and describes our algorithms for generating high quality research trails.
Research trails can be applied in several contexts: as the underlying mechanism for a research task browser, or as feed to an ambient display of history information while searching. A prototype was built to assess the utility of the first option, a research trail browser.