I started programming using the TDD approach.
It seems to work very well especially to avoid initial design mistakes (you actually don't tackle the problem up-front and never design the entire system at the beginning but rather make a sort of "incremental design" laying down many tests, one after the another).
However I found a weird pattern arising that I should avoid (IMHO) when dealing with data parser: an increase need for magic numbers (i.e. mostly field offsets)
Maybe it's just a lack of experience in the refactoring phase?
Aucun commentaire:
Enregistrer un commentaire