Interactive Visualizations to Facilitate Human-Machine Collaboration in Web Browsing (1999)


Description

Browsing and searching are two common techniques for finding information on the World Wide Web (WWW). Though some searching is reasonably well supported by search engines, and many sites provide maps to their contents, browsing has minimal support. In this paper, we describe CoBrowse, a computational assistant intended to collaborate with a user to provide WWW browsing and searching support. It does this by inferring a set of keywords that represent the user's current interests based on the content and sequence of pages visited. With the current interests as a guide, CoBrowse suggests what pages to visit next, displays maps of the WWW neighborhoods around the user's current page, and displays the user's history in terms of the user's previous tasks. In addition to the design and implementation of CoBrowse, we report the results of a preliminary user study that suggests that the support CoBrowse provides can in fact help users find information on the web.