kre at munnari.OZ.AU
Wed Feb 24 17:49:09 UTC 2010
Date: Wed, 24 Feb 2010 08:45:58 -0500
From: "Olson, Arthur David (NIH/NCI) [E]" <olsona at dc37a.nci.nih.gov>
Message-ID: <996D816825CFEA469870126E9050D3F0B68BF10A at NIHMLBX11.nih.gov>
| Can anyone on the list provide a good translation?
That's not me (not even a poor translation), but I wouldn't be
surprised if the report in the e-mail is correct, it is actually
a very intelligent choice for a country that wants to change to/from
summer time at midnight. Most of us (North America, Europe, Aust)
all change in the 01:00-03:00 interval (one way or the other) - that's
partly to make sure the day doesn't jump around.
A rule that had the time changing (especially turning off summer time)
at midnight (or 01:00 wallclock even perhaps) has the potential of having
the day (and even the month in their case) change forward for a microsecond
or something, then flip backwards - even if it never actually enters the
new day (& month) just reaching the trigger point might be enough for
stuff which is scheduelled to happen at midnight at the start of each day,
or the start of the first day of the month, to start running - then start
running again an hour later. In the flip forward case, similar bugs could
prevent the trigger ever happening (the time never cycled from 11:59:59 to
00:00:00, so it is not yet time to ...)
Choosing to move the time back a minute earlier would have no practical effect
at all, to people, the change still happens "at midnight at the end of the
month", but for software, etc, lots of problems could be avoided.
So, if that's what they are doing, very wise I say, and who cares that
everyone else only ever changes to/from summer time at an even hour boundary.
More information about the tz