In actual practice, I regard process and trial/error as two inextricably connected sides of the same scale of knowledge, with the fulcrum of a plausible working hypothesis in between.

Many a great discovery has come by accident while someone was studying something else. ;)