The code looks clean and it works. It's fun to implement your own understandable parser. After doing such an exercise, suddenly a seemingly complex file format doesn't look like such a black box anymore. When I implemented my decompressor, I also looked at https://github.com/pfalcon/pyflate , to understand and verify the inner gzip/deflate RFC workings. Pyflate is also pretty small, especially after removing the bzip2 code, after which only ~450 lines are left. Laiho's code seems more readable, though.