A good principle would be to apply this piece of advice from Hermann Weyl : to understand well a mathematical object, determine and study the structure of its group of automorphisms. I tried to use this method to shed a new light on RH but the structure of the class of L-functions I consider is only conjectural so far.