During the begin of my PhD, I had the goal to develop a wiki software for semi-structured data. The idea was that every page should adhere (maybe in a loose way) to a template. The system had to react on a template edit performing some basic changes on the pages trying to keep them valid. We actually created a simple prototype that used IBM DB2 to store the data in xml files using XQuery as primary query language. The tool was never released mainly because the project was abandoned after realizing that the current technology didn’t let us to create a tool fast enough to guarantee a good user experience.
In the “Reputation system” phase of my research, I worked mainly to understand how to automatically evaluate the users’ reputation in on-line community and how such a reputation can be used to evaluate the users’ generated content. I worked mainly using the Wikipedia dump to analyze the behavior of a big on-line community.