Our implementations are classified by the problems they solve. For each problem, we rate all the implementations according to usefulness on a 1 to 10 scale, with 10 meaning the most useful and 1 least useful. The usefulness of an implementation is supposed to reflect how likely it is that someone looking for code will be happy with what they find. Our ratings are completely subjective and reflect our judgement. Please feel free to send corrections and feedback based on your experiences. Please note that many of these implementations with low usefulness ratings were produced primarily for expository purposes, and can serve as excellent models to study or emulate if not as ready-to-run code.