On Sun, 14 Jun 2015 12:13:31 -0700, Charles Mills wrote: >Leap seconds are irrelevant. Let's say the routine being measured took >86,710 seconds. You would be completely correct in saying it took 1 day, 5 >minutes and 10 seconds even if the run happened to span a midnight at which >a leap second was added. The TOD clock is ignorant of leap seconds. They are >solely about keeping the world's civil clocks aligned with solar noon. > But those civil clocks are used for legal purposes. Failure to account for a leap second might result (possibly; however improbably) in a calculated incorrect date, or even an incorrect year.
All things considered, I believe that UTC was a poor choice for accounting and other aspects of IT. (A smoothed) UT1 would hanve been better. But redefining UTC to remove leap seconds and run free at the TAI rate but 30-odd seconds behind is the worst idea of all. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
