AI

Time, topic and trawl: stories about how we reach our past

Abstract

Legacy web tools attempt to build on information that uses have when they originally conduct web research. In contrast, we examine the information that they have at the time when they attempt to recreate their past. We interviewed 11 non-expert users twice a week for eight weeks in their own physical and computational environments. We used both Google web histories and the prototype Research Trails system as prompts to probe how the participants viewed their past web experiences and how they reconstructed them. The Research Trails system lets users utilize information about both time and topic to help themselves remember and resume everyday research tasks. Based on these observations, a model of users' perceived past web activities informed the iterative refinement of the Research Trails system. The user may see a past action as belonging to multiple categories at the same time or as in different categories at different time