Java and JDBC: tools supporting data-centric business application development

Abstract

Software porting, maintenance, distribution and installation form the bulk of application deployment costs. These high costs have been substantially exacerbated by the move to client server computing. No longer can an application be installed on one system for use by the entire user community. Now it must be installed on all desktops in the enterprise and, at most companies, this will be mean that the software must be ported to and tested on Windows 3.1, Macintosh, Windows95, Windows NT, and often a variety of UNIX systems. The trend towards making business data available to both internal and external users across intranets and the Internet transforms an expensive problem into a completely intractable one. To further complicate the matter, there is an emerging trend in many application domains to move from 2 dimensional windowed user interfaces to high information bandwidth 3 dimensional data visualization and virtual reality systems. These applications are either very difficult or impossible to write using standard GUI builders, generators, or the current breed of portable GUI class libraries. We propose a new architecture for data visualization, discovery, and delivery and argue that the existing development model, and the tools supporting that model, are unaffordable and don't scale to Internet scope delivery numbers.