Introduction

Maria Milosavljevic (1), Fabio Vitali (2), and Carolyn Watters (3)

(1) CSIRO Mathematical and Information Sciences, Australia

(2) University of Bologna, Italy

(3) Dalhousie University, Canada

This volume contains the submissions to the Workshop on Virtual Documents, Hypertext Functionality and the Web, held on May 11, 1999 at the Eighth International World Wide Web Conference (WWW8) in Toronto, Canada.

This workshop was born as the confluence of two previous series of workshops on related topics: the Hypertext Functionalities Workshop series (of which this is the eighth in the series), and the Reuse of Web Information/Flexible Hypertext Workshop series. Fruitful discussions among the organizers of the two parallel workshops at the seventh International World Wide Web Conference (WWW7) in Brisbane, Australia, April 1998, heightened an awareness of the similarities in the topics and the potential synergy of a combined workshop.

By "hypertext functionality" we mean much more than browsing by clicking on "goto" links from one node to another. The focus of the HTF series is on the identification of characteristics that define and describe the "hypertextuality" of software systems. For instance, it aims at describing new ways to view a system's knowledge and processes from a conceptual point of view, to let users access and navigate through the items of interest, to enhance the system's knowledge through comments and relationships, and to customize information and display to the individual users and their tasks. This research is being brought to the forefront within the context of the World Wide Web, and specifically related to virtual documents, as this provides an additional layer of complexity to the issues.

"Virtual documents" are web documents for which the content, nodes or links, or all three, are created as needed. There already exist several kinds of virtual documents on the web for which the content is determined dynamically. First, a template can be used for which node contents are substituted at runtime. Second, applications, like Maple or Mathematica, can be used to generate values for one time use. Third, CGI scripts and search engines can be used to compose virtual documents from fragments of other documents for the user on demand. Fourth, metadata can be generated for summarization for users, where the extraction and summarization is done on the fly for the user. Finally, natural language generation techniques can be employed to dynamically construct virtual documents from underlying data in data or knowledge bases.

The eight short papers collected here provide an overview of related research and reflections on the convergence of hypertext functionality and virtual documents on the world wide web. We will give here a brief introduction to these works:

At an abstract level, Watters and Shepherd provide a list of fundamental, systemic issues to be identified when discussing the production and use of virtual documents in the world wide web, and Crowston and Williams discuss the value of genres in Web documents, structures that are also important even in a new medium such as the Web. At a development level, Green, Milosavljevic, Dale and Paris report on their actual experiences with the development of systems generating web documents on the fly, especially with the aid of natural language processors.

Providing formal grounds to the issue of dynamically creating virtual documents, Azevedo de Lima and Price propose a methodology using design patterns, which contrasts nicely with the model of Ranwez and Crampes based on composable Information Bricks.

At a methodological level, several papers identify the atomic blocks of virtual documents and discuss how to enrich them with the information needed to generate different virtual documents. Martin and Eklund propose to embed machine-understandable information (conceptual commands) in the source, real documents. Rock, Cawsey, McAndrew and Bental rely on associating documents with metadata described using standard metainformation sets. Finally Caumanns proposes an architectural framework for software modules to create, collect and compose chunks of inforation into virtual documents.

We believe these presentations and the discussion following in the workshop may help in answering important questions related to virtual documents and virtual application domains. How are virtual documents defined and managed? The management of this class of documents requires new understandings of bookmarking, versioning, authentication, structure, ownership, navigation, collaboration, and reuse of components. Issues of security, data protection, verification, and access control need to be addressed. Finally we need to address questions about how to determine if these web information systems are actually improving service to the users.

Of course, this workshop is part of a growing body of connected events on similar topics, which is worthwhile to list here: