Web application development : a study on UML Web Application Extension

• Main Authors: Norberg, Tobias, Oskarsson, Andreas, Kling, Martin
• Format: Others
• Language: English
• Published: Blekinge Tekniska Högskola, Institutionen för programvaruteknik och datavetenskap 2002
• Subjects: Webb; Internet; Modellering; UML; WAE; Web application; Computer Sciences; Datavetenskap (datalogi); Software Engineering; Programvaruteknik
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:bth-1689

The complexity of Web sites are increasing and transforming into Web applications that contain business logic, interactivity, transaction handling and states. This phenomenon forces the Web developers to adapt more traditional software engineering techniques to keep the Web applications error free, maintainable, reusable, well documented etc. Many Web developers do not use any engineering techniques at all and design mainly to create as fashionable applications as possible with no regards on the application?s functionality. This results in applications that are hard to maintain and with poor functionality. The purpose with this thesis was to see if the use of a more traditional software engineering technique, namely the Unified Modeling Language with the newly added Web Application Extension, resulted in a Web application with good design regarding the maintainability of the application. To investigate the maintainability of an application, the maintainability was further divided into three sub criteria: extensibility, reusability and documentation. These three criteria were then applied on a case study were a Web application was designed. From the analysis of the final design, using the three criteria, the maintainability was derived. The result of the entire investigation showed that the UML WAE had a good support for extensibility, fair support for reusability and very good support for documentation. From these results the main conclusion was derived, that the use of UML WAE resulted in good design regarding the maintainability. However, the result is limited to our case study and the design created in that case. The result may have been different if the three criteria had been applied on a different case. Another aspect to consider is that the quality of a design is often dependent on the knowledge of the persons that carry out the design.