[LEAPSECS] The Debate over UTC and Leap Seconds

[ashtongj]

On 2010-08-10 11:03 AM, Poul-Henning Kamp wrote:

> The much larger body of non-working code, which needs to be fixed

> if we continue the leap-second stunt, is not mentioned once in the

> "Favouring change" column.

>


If one actually wants to compute legally correct durations given the 
starting and ending date and time, written in a legally acceptable 
manner, one has a great many programs to fix, and in some countries and 
time periods it is impossible to decide the right answer. This is 
because many countries legally observe UTC-based time, and the leap 
seconds must be counted in the duration. Many countries nominally 
observe GMT-based time, but their national time laboratories provide 
UTC, so it is hard to decide what the actual legal time standard is in 
those countries. These problems will continue to exist even if future 
leap seconds are eliminated, because the duration to be calculated may 
be in the leap second era, or may have begun in the leap second era.

Fortunately this problem does not come up very often, but there are many 
computer functions that purport to do this but really don't.

Gerry Ashton