[LEAPSECS] Crunching Bulletin B numbers (POSIX time)

[Warner Losh]

On 02/22/2011 09:33, Tony Finch wrote:

> On Mon, 21 Feb 2011, Ian Batten wrote:

>> But the ITU proposal is akin to abolishing leap years and pretending

>> that there is no gregorian correction...

> Not really. It's more like switching from an observational lunisolar

> calendar to an arithmetic calendar.


Right now the function for determining a leap second is effectively:

     if (date == june 30 || date == dec 31) && random() < 0.33)
         return true
     else
         return false;

while the leap-year code is:

     return (year % 400) == 0 || (year % 4 == 0 && year %100 != 0)

which will last us about a thousand years.

Warner