3 ms·
Daniel J. Bernstein dissects and disputes epidemiology paper modeling
- ncmncm 7y agoI have talked with scientists about the risks of bugs in dodgy software written by inexperienced grad students, or themselves. They have uniformly considered it entirely a non-problem. Their reasoning was that either the results would be replicated (good!) or wouldn't be (fine!). Either way, nothing to waste time worrying or doing anything about. It did not lead me toward confidence in the reported analytical results of experiments.