[LEAPSECS] Meeting with Wayne Whyte

[Stephen Colebourne]

On 1 February 2011 05:12, Mark Calabretta <mcalabre at atnf.csiro.au> wrote:

>>It is also a central problem of time_t: how do you map this

>>non-uniform-radix notation onto a uniform count that must always satisfy

>>properties that explicitly mandate a uniform-radix.

>

> Vide the mapping of calendar date to Julian Date.

>

> The fundamental problem is that there is no formula for determining

> when leap seconds occur.


No, the rules for inserting leap seconds are simple enough in
computation terms. The problem is entirely with the 23:59:60 notation
being unacceptable to the API design.

API users call getSecondOfMinute() and require a value 0-59. That UTC
doesn't do that is a unfortunate reality, but can be handled by
smoothing, provided a 2nd API is available (not the main one) that
allows a determination of the 0-60 value, eg.
getSecondOfMinuteWithPossibleLeapSecond() - JSR-310 effectively puts
that on a separate class - UTCInstant, expressing it as getNanoOfDay()
which may reach the equivalent of the 61st second.

Stephen