o The code must be understood by the compiler

o Avoid (if possible) warnings due to generics and the use of raw types

o Avoid every warnings (if possible) except the stupid ones

o Anybody can reformat the code, but only into the Sun standard coding style

o Only write code that is usefull :-)

Datatype usage:

o Use 'double' for every real numbers. Only exception is float[] for samples (or more generally for large arrays)

Naming conventions:

o Instance variables holding input ports are named inXXX. Output ports: outXXX

o Integer multiplier: "factor", floating point multiplier: "ratio"

