[LEAPSECS] What's the point?
sla at ucolick.org
Wed Feb 9 10:09:28 EST 2011
On Wed 2011-02-09T08:56:41 +0000, Clive D.W. Feather hath writ:
> Firstly, people don't use zoneinfo for that sort of thing. You'll end up
> with millions of confused users trying to work out why there's a 30 second
> difference between one timestamp and another.
This exchange makes it clear that some people are already doing that.
It makes sense if your site is so leap-sensitive that you've chosen
something like the Meinberg NTP master which can be asked to emit TAI
instead of UTC. It doesn't make sense for most other folks. (And
the remark there that seems to be from the QNAP guy is, quite frankly,
scary if it represents the common level of understanding of zoneinfo.)
> Then what about the "rubber second" period of history. From memory, the Olsen
> database only goes to a precision of 1 second.
Fixing the past with the "right" zones has already been tried; it
fails in ways that should never be repeated unless your site never
interacts with a POSIX-compliant site.
The past of UTC is broken, unfixably, and that's part of why the ITU-R
should shun the name UTC if they want to deconfuse precision
timestamps for the sake of the future.
Steve Allen <sla at ucolick.org> WGS-84 (GPS)
UCO/Lick Observatory Natural Sciences II, Room 165 Lat +36.99855
University of California Voice: +1 831 459 3046 Lng -122.06015
Santa Cruz, CA 95064 http://www.ucolick.org/~sla/ Hgt +250 m
More information about the LEAPSECS