The problem is with the media and the way things were spun. Correcting that should not involve stifling the workflows scientists use. The GitHub analogy is a good one as what you're suggesting is the equivalent of only pushing perfect code.