On Dec 8, 10:04 am, Shane Geiger <[EMAIL PROTECTED]> wrote: > What is so obvious about dealing with months that vary in length and the > leap-year issue? Nothing. If you were born on a day that does not > exist every year (Feb 29th), how old are you on Feb 28th?
X years, 11 months, 28 days or Mar 1 of > non-leap years? X' years, 0 month, 1 day If you were born on Feb 29th, then you would be one > month old on March 29th, but would you be one year, one month and one > day old on March 29th of the next year? or would you merely be one year > and one month old? 1 year, 1 month, 0 day ; why would there be one day more ? People born on the 28th would be one year, one month and one day old. If two dates have the same day-in-the-month then the difference is X years, Y months and 0 day I understand that there is no possible conversion from a number of days to a (X,Y,Z) tuple of (years,months,days), and the reverse. But the difference between 2 dates can be unambiguously expressed as (X,Y,Z), and given a start date and an interval (X,Y,Z) you can also find the end date unambiguously, provided the arguments are valid (for instance, 1 month after the 30th of January is not valid) Regards, Pierre -- http://mail.python.org/mailman/listinfo/python-list
