Posted by Gordon Burditt on 03/08/07 00:49

>> By the way: please use GMT+0 (or unix timestamps) to keep dates on a
>> database. If your country follows DST and you use the local date, you may
>> run into lots of problems.
>
>Not just "may" - it's a slam-dunk!

Maybe, maybe not. It depends on what you're using the time stamps
for. And *future* times are a problem, since your conversion from
human local time units to UTC may turn out to be inaccurate if
governments screw things up. I'll agree that having the OS clock
in UTC is a good idea. (Let's see, we'll count heartbeats while 1
minute goes by, starting at Sunday, 1:59:30 Mar 11, 2007. Wait!
This guy's pulse rate is 3600! Yikes!)

If I've got a clock that counts seconds and is expected to be accurate
for the next 50 years, you'll need to fiddle with the display algorithm
however many times governments screw with daylight savings time. This
corresponds to the OS daylight savings time patches.

If I'm a stay-at-home type of guy (don't change time zones much)
if I've got an appointment calendar with a lunch appointment every
Monday at noon, chances are much better the program will do what I
mean if it stores that future appointment in local time. Chances are I
mean the appointment is at noon, local time, regardless of whether
it's a multiple of 7*24 hours since the last appointment. I might
even mean noon, whatever local time zone I am in at the moment.
This corresponds to the Microsoft Exchange Server and Outlook daylight
savings time patches.

If I do change time zones a lot, I probably want to record the time zone
of the locality to which the appointment applies. Few users who need
this ability will really use it to advantage until they get burned.
And what, exactly, is the "locality" of a phone conference with members
in several time zones?

>Microsoft is announcing this week that you WILL run into
>problems on their systems... as are several other companies.
>
>In 2005, the Republican congress decided to mess with DST in
>order to trick some particularly foolish voters into believing
>that it will save energy dollars.

Has anyone come up with proof that this will save more energy than
changing DST in the *other* direction? Most of the offered
explanations (mostly by news media) involve DST actually changing
the number of hours the sun is shining.

[Back to original message]