1 Collaborative Categorization on the Web: Approach, Prototype, and Experience Report François Bry and Holger Wagner Institute for Computer Science, University of Munich Oettingenstr. 67D-81538 München, Germany http://www.pms.informatik.uni-muenchen.de/ September 2003 Abstract Collaborative categorization is an emerging direction for research and innovative applications. Arguably, collaborative categorization on the Web is an especially promising emerging form of collaborative Web systems because of both, the widespread use of the conventional Web and the emergence of the Semantic Web providing with more semantic information on Web data. This paper discusses this issue and proposes two approaches: collaborative categorization via category merging and collaborative categorization proper. The main advantage of the first approach is that it can be rather easily realized and implemented using existing systems such as Web browsers and mail clients. A prototype system for collaborative Web usage that uses category merging for collaborative categorization is described and the results of field experiments using it are reported. The second approach, called collaborative categorization proper, however, is more general and scales better. The data structure and user interface aspects of an approach to collaborative categorization proper are discussed. 1 Introduction The World Wide Web is a rich and complex space for retrieving all kinds of information in academic and commercial contexts. Many people are already collaborating in the effort to use this medium efficiently, but doing so without dedicated technical support. Instead, people are sending URIs pointing to documents they find interesting via email, with annotations that are lost when the email is deleted. Bookmark collections are manually converted to Web pages and uploaded to the Web - but peers must still be informed about the location of such pages and maintaining them is cumbersome. Some Web pages offer guest books or discussion forums, that are then used by an emerging community of people interested in the contents of such pages - but lacking an integrated system, their users have to cope with different user interfaces, no interaction between the discussion forums and generally no unified experience across different platforms. The present paper is based on the ongoing development of an integrated approach to collaborative Web usage - supporting these forms of collaboration with a software system that integrates features existing distributed among various applications into a consistent concept. In [Wagner2002], a terminology has been laid out, defining the terms relevant to the subject matter. Concepts relevant to an integrated approach to