"Normally, science advances by trial and error. When an experiment fails, researchers question assumptions, formulate new ideas and then design better studies."
1/ it is not normally but most frequently
2/ when there is no experiment there is no facts, an experiment should be redone to assess the robustness of results
3/ facts allow human action to be directed toward confident results but uncertainty is always present
4/ scientific researchers developed methods to reduce uncertainty about interpretation of facts
5/ this uncertainty is easier to reduce in physics as for instance the probability of a bridge to fall down than in biology where individuals are genetically different and interact intensively with the milieu
6/ Science people are aware that theories paradigms and conceptions improved only by controversy and are submitted to falsifiability