Still to be done:
* compute(u32,u32)
* bugfixes, increase efficiency with doing the modulo (i.e. avoid
  the intermediate call in the chain m/n -> n/m -> n%m/m)