That doesn't work. 13 does not divide evenly into the 365.24 rotations the Earth performs as it revolves around the Sun once. Nor does 28 divide evenly into the 27 days and 8 hours that elapse for every cycle of the moon.
We obviously need to move a large number of massive weights down near the equator to slow the earth down till we achieve a nice round number of days per year. Say 256. Each year the International Earth Rotation Service http://www.iers.org/iers/ could issue a report showing how well the tuning was working and if weights needed to be added or removed. ;-)
More seriously, I was just struggling with leap seconds this morning, trying to reconcile un*x time, GPS time, UTC time and TAI time. What is the current feeling in the linux community with respect to leap seconds? Does the kernel's time still jump around when a leap second is applied or subtracted or can the corrections be applied in user-space like the DST corrections are?
-wolfgang