Kitagawa, Hiroyuki, Morishima, Atsuyuki, Kambayashi, Yahiko, Kitsuregawa, Masaru, Makinouchi, Akifumi, Uemura, Shunsuke, Tanaka, Katsumi, and Masunaga, Yoshifumi
With the broad acceptance of the World Wide Web, the Web has been widely used for publishing and disseminating information originally stored in various information sources. They include databases, document repositories, and Web servers. In constructing Web structures on top of the heterogeneous information sources, data acquisition is an essential issue. Also, design of data layout on Web pages is another important issue. A lot of work has been reported on the problem so far, and some tools and systems are used in practical applications. In this article, we survey the current approaches to this problem in the first part, and show our approach in the second part. The current approaches to the problem can be classified into the program development approach, the template-based approach, and the mediation approach. In the first approach, application programs to acquire data and to generate Web pages are developed from scratch, and they are invoked through CGI, Servlet, and/or Web server API. In the template-based approach, some kinds of layout templates with scripts and queries are used to specify the data acquisition and data layout simultaneously. In the mediation approach, a common mediation data model is introduced to specify layout-independent logical data acquisition or data integration. In this article, we mainly explore the mediation approach, since this approach is most promising in the context of large-scale Web structure construction. We give an overview of the current technology related to this approach from a number of important design viewpoints. The second part of this article explains some details of our approach. It takes the mediation approach. The common data model is based on nested relations and ADTs, and interactive visual authoring facilities are provided for data layout specification. [ABSTRACT FROM AUTHOR]