4 WWW Needs More Interaction Hermann Maurer IICM, Graz University of Technology hmaurer@iicm.edu Abstract. In this talk we argue that WWW is seen much too much as an information system with a very limited variety of possible interactive features. We claim that some of the most important possible functions have been neglected, sofar. 1 Introduction WWW (or the Web, as we will call it henceforth) was originally conceived as a distributed information system, initially to give access to new physics reports, world-wide. Although the Web has expanded its functionality in many different directions some of the most obvious ones have been still neglected. It is true that forms for feedback and necessary for transactions have been introduced and do play a major role by now (e.g. in e-commerce), and that email, chat, Web telephony and Web video conferencing have been much widening possible applications of the Web. Yet, a number of equally important aspects have been largely ignored. We will describe two of them, customization of information and tools for cooperation in some detail in what follows. 2 Customization of Information In most Web systems the Webmaster is a "dictator:" the Webmaster defines what and how information is to be seen, i.e. is structured. It is curious that this has not caused a rebellion already years ago: is it not ludicrous that the head of a company when accessing the firm's server sees the same information structured in the same way as the head of research, the head of personnel, ordinary employees and maybe even customers? And clearly the same is true in school-, university- or governmental-environments. Thus, it should be clear that the type of material that can be accessed should be definable by some authorization mechanism (typically password based) but, much more, the structuring of the material has to also depend on the user. To put it into modern terminology: documents or groups of documents must be accessible by complex ontologies, i.e. providing different "views" or "access paths" depending on the user group. Hence, a first step in this direction is to allow users to identify themselves (with the simple http protocol of the Web this requires unfortunately strange work-arounds such as cookies or session keys) and, depending which group they K. Beiersdörfer et al. (eds.), Informatik’99 © Springer-Verlag Berlin Heidelberg 1999