Open science, reproducibility, coding errors