in reply to Re^3: Efficient enumeration of pandigital fractions
in thread Efficient enumeration of pandigital fractions

Why not base-2 ... could you elaborate?

8 * 1234 (the smallest possible denominator) = 9872, which is just 4 away from the largest possible numerator, 9876, so there isn't much wiggle room there. But 8 is already taken, and the next possible combination 9765, is not divisible with 8, and the quotient is less than 1234.

In general, zyxw...(base/2-1) - (base-2) * 1234...(base/2-1) = base/2 - 1.