OK, you claim that I gave you not a value. I claim that I gave you a correct value, and a correct procedure to validate the claimed error. Moreover, I claim that if you insist that for any given value, no matter how created, the expression ROUNDDOWN(x) - x may produce a value greater than 0, and that is not because of the Rounddown, but because of subtraction, then it is even more severe problem, not a mere correction introduced for very clear and sound reasons, but a plain violation of all fundamental arithmetic rules and also of IEEE 754 requirements, which means that the claimed introduction of the FP error by subtraction canât be its cause. Then I claim that all this shows that you simply donât understand what you are saying, or alternatively that you are a simple troll.
One can validate that the values 1.99999999999999, 1.99999999999999 + 5e-15, 1.99999999999999 + 10e-15 form a correct sequence, and their subtraction gives expected results. Further, subtracting the last value from the result of rounddown of the second last gives 0. And there is a UI limitation imposed by Excel, not using and not showing more than 15 significant digits in literals, but no restriction in the values created by calculation.
