Content area
Full text
The technology behind distributed processing, frameworks, and the Java language environment is relatively new and changing fast. This requires that current application development tools adapt more quickly to requirements of their users. This paper describes how these tools can be made more flexible and customizable. As an example, the architecture of Business Component Prototypes for IBM SanFrancisco(TM) is presented. In its current form, Business Component Prototypes is a tool to develop prototype applications provided with IBM SanFrancisco v 1.4. Its objective is to help new SanFrancisco developers create small prototypes using the SanFrancisco foundation layer programming model, without requiring the heavy tool set used for production application development. IBM SanFrancisco is a Java(TM)-based set of components that allows developers to assemble server-side business applications from existing parts, rather than build "from scratch."
The use of object technology for application development is slowly becoming mainstream. Despite this, the promise the technology seemed to hold-an increase in productivity by an order of magnitude-has not yet been fulfilled, as we demonstrate later.
We start to recognize that more is needed than just object-oriented languages and object-oriented analysis methods. Standardization, ready-to-use components, and development environments where major parts of the code are generated for the developer should solve the problems. Standardization is easier if all developers use the same programming language. A large segment of the information technology community has decided that the Java** programming language and specifications for JavaBeans** and Enterprise JavaBeans** will bring this much needed unity. As a communication protocol for documents and structured data, xmL (Extensible Markup Language) is emerging. Based on these standards, serious component frameworks are starting to appear, such as IBM'S SanFrancisco*.
Are components the solution?
As a pioneer of component technology, Brad Cox' should be mentioned as the inventor of the Software-- IC**. It seems that more than ten years after he published his seminal book, his ideas finally have become a reality.
David Taylor2 has an interesting argument why CASE (computer-assisted software engineering) tools are not the solution. Assume a traditional development effort where equal development time is spent on analysis, design, code, test, and maintenance. Consider then the perfect object-oriented CASE tool, which would make analysis and design 30 percent more effective, reduce implementation to zero, and...