Many years ago I ran into this paper "Self-indexing inverted files for fast text retrieval" http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.18.8282&rep

Many years ago I ran into this paper "Self-indexing inverted files for fast text retrieval" http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.18.8282&rep=rep1&type=pdf http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.18.... It's short and to the point. And then I implemented all that ... in PHP and MySQL :) It feels daunting at first, but once you understand what it wants you to do, it's actually not that hard (for this particular paper, and this particular approach). However, you do want to employ a stemming library to normalize word forms.