Linux "date" command ignores leap-seconds?

[Christopher Browne]

On Thu, Jan 17, 2013 at 4:26 PM, D. Hugh Redelmeier <hugh-pmF8o41NoarQT0dZR+AlfA at public.gmane.org> wrote:
> | From: Eric B <gyre-Ja3L+HSX0kI at public.gmane.org>
>
> | On Wed, 16 Jan 2013, David Collier-Brown wrote:
>
> Of course only you saw this since the list is still censoring DCB.  I
> presume that he CCed you.
>
> | Adjusting the epoch seems ugly; it should be a fixed point in
> | time.
>
> I absolutely agree.
>
> So when should the epoch start?  23 October 4004BC?  Do we know the date of
> creation, to the second?  Of course POSIX says the Epock starts in
> Nixon's reign.  (Who remembers the Nixon Switch in UNIX?)
>
> Once UNIX systems get into relatively different frames of reference,
> we'll have to rethink this.  Time starts to look like GIT versions
> with branches and awkward merges.

I'm pretty happy with the notion of the epoch starting in 1970.

There's a fairly easy remapping between POSIX timestamps and strict
UTC; one can essentially step ahead by 6 month increments and see
at which times leap second adjustments took place.

In effect, I'd prefer that POSIX be "strictly seconds since 1970, without
adjustments", and then use that as the basis for computing UTC or UT1.

In effect, this is an equivalent to Julian Days, which are a strict "we
keep counting" measure of days since Jan 1 4713 BC.

I actually would suppose that I'd prefer COBOL / "ANSI Days" to
Julian Days, as the epoch for that is Jan 1, 1601, which is conveniently
nearer in history, with somewhat rounder numbers.
-- 
When confronted by a difficult problem, solve it by reducing it to the
question, "How would the Lone Ranger handle this?"
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists