pick the low-hanging fruit
time is the most valueable
assume people are good
programming: assume that the best possible future system is in place (find better quote by that guy) don't prematurely optimize