Framework for web application domain knowledge extraction

A decade ago a web application e-Student was built with aim to provide electronic support for student enrolment and examination/alumni records management at the University of Ljubljana. Due to issues emerging from the Bologna reform a new e-Student is to be build using a modern technology in the near future. The old e-Student encapsulates a huge amount of domain knowledge. Unfortunately, it was developed using agile approach resulting in poor technical documentation, thus an alternative approach for the domain knowledge extraction has to be defined. In the paper a framework for an effective web application domain knowledge extraction is defined. It has five elements. The main principles (1) of extraction are defined to perform effective reengineering of different application views at a defined abstract level. A proper knowledge representation using diverse models (2) has to be determined next, and the Model Driven Architecture using UML models is considered a suitable choice. The procedure (3) for extraction has to be defined using appropriate (usually custom made) tools (4) and performed by skilled staff (5), possibly members of the old development team. The use of framework is demonstrated on the web application e-Student outlining several custom made tools, the results and the most valuable lessons learnt.