Research On Database Semantic Discovery Method Based On Web Application

In recent years, with the rapid development of software technology, the scale andcomplexity of software system has been increasing, so a large part of softwaredevelopment work has focused on the maintenance and renovation of existingsoftware systems. Database-centric legacy systems, such as information managementsystem, usually relied on a large number of database tables. Due to the lack of complete andclear documentation, database maintenance personnel were difficult to understand databasedesign, and they couldn’t accurately grasp the semantics of the field in the database table. Indatabase-centric web applications, program code would be a rich source of semanticinformation about the database fields. However, there were few of research ondatabase-centric web applications to find the semantic of field.To solve this problem, a model has been proposed to discover the database fieldssemantic in the database-centric web application. Firstly, in the model the HTML code onthe web page was analyzed to extract control names and its description by using webinformation extraction technology. Then in the model the source code was analyzed byusing program understanding technology, which converted the source code into systemdependence graph and then analyzed the data flow of control variable to find therelationship of control and field. Finally, the semantic of field would be found by using theresults of two steps above indirectly.Finally, the proposed method was validated by experiments, which proved the methodwas correct and effective by using precision and recall, and the future direction of furtherresearch and improvement were also discussed.