Browser Interface

In the early days of web browsing, the "Bookmarks" feature of web

browsers looked very useful. Over time, however, an active users

accumulated thousands of bookmarks and the feature had limited utility

at best. A number of services exist [1] that can store bookmarks on a

web site and even facilitate sharing with other people, but none of

them solve the fundamental user interface problem.

We propose to address this problem by developing a browsing interface

that provides multiple views of a bookmark file: some could be

relatively simple such as ordering by date, server, or topic (as stored

in document metadata) but will also include unsupervised clustering

based on document contents [2], insertion time [3-4] as well as

hyperlinks and individual and collective usage information.

I imagine two possible scenarios for this project: one is a "My arXiv"

feature for the arXiv e-print library which will let patrons store a

list of their favorite papers in a convenient location. (This will

entice users to store information about their reading preferences which

will make it possible to make further developments towards collaborative

filtering.) Another possibility is the development of a Mozilla

extension that works with a server application for automatically

categorizing web bookmarks.

Contact: Paul Houle ()

Credit Hours: 3-6 (negotiable)