That's why my main point was that scientists need to be taught that code is as likely a source of errors as anything else in the scientific process.