Topic tree

Abstract

Web History is literally a gold mine in terms of useful information for User Modeling. When analyzed over a protracted period of time it gives a detailed breakdown of the user's interests; the relationship between short term interests and long term interests, how one transitions into the other, and vice versa. Statistics from [15] show that anything between 58% and 81 % of all web browsing consists of reviewing content and revisiting pages and reviewing similar content. This, of course includes regular visits to forums and blogs and portals, whose content changes very often, Social Networking sites and sites dedicated to upcoming events. Yet by looking at how the most popular browsers handle history, you wouldn't be mistaken in thinking that someone (either the statistics or the browsers) got it wrong somewhere. Current history functions are very poor visualization aides, with the leading 'off the shelf browser in terms of visualization probably being Apple's Safari (which uses images to make history more memory friendly) still has miles to go in terms of organization. Safari (off the shelf version) offers the most, history that is both searchable, sortable and has decent visualization. And still, it will just present you with something in a list format that can only be navigated linearly. Browsing history is just an exercise in jogging memory, and a linear structure is just not very good at it, being 'stuck' at a single level of abstraction. The information stored doesn't provide very suitable guidance: page title (which might have been automatically generated, and not so helpful), an excerpt, and the last time you visited.