Should we (ask authors to) open (their) data?