We evaluate the efficiency and effectiveness of each
approach by running over a TREC sized test collection
and comparing searching times and relevancy.
We run experiments with out-of-the-box settings
and also try to do language specific improvements for
both relational databases and IR Libraries.