- Check why some errors seem to be not correctly added. - Check and optimize rt_fix_min_iwl() (might rename it to rt_fix_norm). - optimize basic functions add, sub, mul, div (might implement them using inline assembler). - Check maximum size of arguments for log2(), log(), sin(), cos(), ... - document maximum error, which can occurr in different functions - find out why the result of rt_fix_exp(-22) and rt_fix_log(1) is -0.000 (where comes the sign from - I don't want it there.) - be conforming to ISO/IEC 10967-1 (Information technology - Language independent arithmetic)