Text this: Error and the growth of experimental knowledge