We have implemented a few small changes to the transforming and loading the rdf triple store:
1.streamline upper/lower cases, eliminate special characters in the URIs;
2.breakdown habitat locations and flowering time;
The search logic is more streamlined and took less time. The performance issue under current set up (windows xp virtual machine with apache+tomcat+mysql5.1 ) is inconclusive, we will have to test on a real server to test and proceed.
The front end includes basic wire frame and core elements. Done entirely using Dojo controls and Ajax request.
Path forward: The emphasis will be on the search logic such as query result estimation, Sparql performance and etc. The front end will be added with more features such as result number limit and truncate, user options etc. If the core part is well received, we'll work out a feature list and add on to the project.