the point is not that subtracting 20 gives you the ability to do math, but that POSIX seconds are not equal to real seconds. a TAI second is equivalant to a standard second, A posix second is slightly longer than a normal second. The fact that the proposal specifices that ClockTime's are in picoseconds AND that the clocktime may be a POSIX timestamp are in direct violation with each other.

Posix time_t's should not be thought of as an 'approximation' of real time or tai time, they have a well defined meaning and can accuratly represent specific times with slightly less than a second of accuracy. Treating them as representing standard seconds however is clearly incorrect, as they do not. they represent a number that can be trivially converted to an exact UTC time.

Ok, I see what you're saying. Could you point to somewhere on the web that describes this definition of POSIX time_t? I looked in POSIX 1003.1-2003, and it has a different definition (the one I quoted earlier). http://www.opengroup.org/onlinepubs/007904975/basedefs/xbd_chap04.html#t ag_04_14 Indeed, the system I'm using here doesn't correspond to your definition of POSIX time_t, it uses the definition from the link above. To clarify, under this definition of time_t, the clock ticks at second intervals, but time_t cannot represent certain times. Try the following code in GHC: -------------------- import System.Time date1 = CalendarTime {ctYear = 1998, ctMonth = December, ctDay = 31, ctHour = 23, ctMin = 59, ctSec = 59, ctPicosec = 0, ctTZName = "GMT", ctTZ = 0, ctIsDST = False} main = do print (toClockTime date1) print (case toClockTime date1 of TOD i j -> TOD (i+1) j) -------------------- The program constructs a CalendarTime at the last point a leap second occurred, converts it to time_t, adds one, and converts it back to CalendarTime again. Let's see what happens: ~/scratch > TZ=UTC ./timetest Thu Dec 31 23:59:59 UTC 1998 Fri Jan 1 00:00:00 UTC 1999 Where's the leap second gone? There are *two* real seconds between these two times, yet I only added one to the time_t value. ~/scratch > TZ=right/UTC ./timetest Thu Dec 31 23:59:59 UTC 1998 Thu Dec 31 23:59:60 UTC 1998 Aha! Setting TZ to "right/UTC" does the trick: now I can see the leap second, because time_t includes leap seconds. Cheers, Simon