Occasionally while reverse engineering compression schemes I run across techniques I don't recognize even after constructing a working re-implementation in C. Exotic algorithms or common ones obfuscated by optimized implementation.
I would like to identify those algorithms so that I can study the theory behind them. Maybe some of the compression enthusiasts here are interested to take a look sometimes?
For example, here is something I worked on recently:
http://asmodean.reverse.net/pages/expcf.html
The first stage of decompression (unwtf() in that code) reads a table of character weights(?) and then fills a string replicating each character by the weight. Then it loops doing some divisions to compute indexes that I don't understand enough to describe in words.