That's odd for a business so concerned about selling to academia. You'd expect journals to reject papers where the number-crunching went through a "black box" like that.