Correct. But since the earthly instant might have occured anywhere on the
planet we have a span of 3 days (+- 1). While any event measured in minutes
can have occured on up to two different local time zone days if we don't
have the measure of time nor even the location, we don't know if the event
could have be reported by another location on the planet on the previous
calendar day or the next.
Imagine a date/time reported in a database as 2010-12-12T01:00 without
knowledge of the "local time" where reported.. If it was Baker Island.. If
it was Kiritimati.... Searching with a date/time specified in UTC in such a
database.. What is the match? What is the precision?
On Mon, 23 Nov 2015 20:35:54 +0000, Byrd, Donald A. wrote
> There's only one International Date Line; time zones on the earth span
just 24 hours. So it seems to me an instant can't be on more than two
different days, regardless of where and when it occurs. Munich's time zone
is nine (I think) hours west of the Date Line, so an event at, say, 11:45
PM on Nov. 30 there is on December 1 anywhere more than a few hundred
kilometers (or some such) to its east, up to the date line. It's 15 hours
from the Date Line in the other direction, so the earliest it can be in any
time zone is 15 hours earlier, i.e., 8:45 AM on Nov. 30.
>
> Furthermore, an instant also can't be on _less_ than two different days,
regardless of where and when it occurs. Well, it could be argued that an
instant at exactly midnight somewhere is an exception! Anyway, for
practical purposes, every instantaneous event occurs on exactly two
different days.
>
> Am I thinking clearly?
>
> --Don
>
> On Nov 19, 2015, at 5:13 PM, "Denenberg, Ray" <[log in to unmask]> wrote:
>
> > So Neil [WINDOWS-1252?]Armstrong’s moon event is nominally 1969-07-20,
[WINDOWS-1252?]that’s the [WINDOWS-1252?]“official” date that it occurred
(nevermind that it was on the moon, [WINDOWS-1252?]let’s just say July 20
is the [WINDOWS-1252?]“real” date, for discussion sake). Some observers may
have recorded it as July 19. And I suppose somewhere else it was recorded
as July 21. Would you call this [WINDOWS-1252?]“two-day” precision?
Three-day precision?
> >
> > An event in Munich may occur close to midnight so it is the next day in
Samoa (and the previous day somewhere else?) But it might be November 30
in Munich, so it is December in Samoa, or it might be December 31 in Munich
so it is the next year in Samoa. So the first case is 2 (or 3) day
precision, the second is month precision and the third year precision. And
maybe it occurred 2009, so there you have decade precision.
> >
> > Still, all these examples are based on cases where the estimate is no
more than a day off. I [WINDOWS-1252?]don’t think I can convince the ISO
people based on that. Are there other examples?
> >
> > Ray
> >
> >> -----Original Message-----
> >> From: Discussion of the Developing Date/Time Standards
> >>
> >> [---- SNIP ----]
>
> ---
> Donald Byrd
> Woodrow Wilson Indiana Teaching Fellow
> Adjunct Associate Professor of Informatics
> Visiting Scientist, Research Technologies
> Indiana University Bloomington
--
Edward C. Zimmermann, NONMONOTONIC LAB
|