Are online comments 'bad for science'?