On 7/30/2026 12:00 PM, Brooks Harris via tz wrote:
I do think ''permanent standard time' would be the better approach to
ending DST for a few reasons; A) it would put us in sync with the
majority of humanity who do not observe DST, B) it results in the most
natural daylight hours, C) it avoids the "kids in the dark" problem,
and D) it's much easier for TzDb and downstream to implement.
If the US ends seasonal DST we will be a country that does not observe
DST. Not observing DST does not mean re-adjusting the clocks back to a
degrees from solar noon standard. For example, Saskatchewan Canada "has
not observed DST" since 1960 but they set their clocks forward before
stopping the changes. This year British Columbia, Alberta and Northwest
Territories decided to follow and stop observing seasonal DST after
setting forward their clock. Yukon Canada set their clocks forward and
"locked the clock" twice. They started out as -09 and are now at
-07. What daylight hours are "natural" is an opinion. If one looks at a
graphic showing the offset of "solar noon" from "clock face noon" you
will see more of the world letting solar noon occur closer to 1pm or 2pm
clock face time than at 11am. [1] The "kids in the dark" could be
applied to morning darkness or late afternoon darkness. Arguably people
are better drivers in the afternoon (unless they are rushing around
trying to get something done before an early sunset?).
It is much more convenient to work with full hour offsets than a system
where clocks could be minutes apart due to arbitrary obedience to a
"solar noon". Imagine a "9am" morning meeting that would occur halfway
between sunrise and solar noon every day. If that meeting involved other
cities one would need to decide which city's sunrise and solar noon
would be followed. The current time zones do that, breaking away from a
sunrise/sunset based standard and following a fraction of a rotation
standard (an hour is 1/24 of a rotation, a minute is 1/1440 of a
rotation, etc.). It is also a lot more convenient for engineers who do
not have to design a clock that would follow the drift of solar noon
throughout the year. Define sunrise as 6am every day and sunset as 6pm
every day and divide that in to "hours" and "minutes" then build a
timepiece that is accurate? Modern time zones are a lot easier to implement.
Indeed TzDb and downstream should investigate implementing H.R. 139.
I'm in favor of Paul's approach changing FORMAT to the zone's offset.
For America/New_York (using 2027 as start date):
# Zone NAME STDOFF RULES FORMAT [UNTIL]
Zone America/New_York -4:56:02 - LMT 1883 Nov 18 12:03:58
....
-5:00 US E%sT 2027 Mar 14 2:00
-4:00 - -04
I think possibly reinstating FORMAT as "EST" later could be
problematic and better to just stay with the FORMAT as the offset.
I'm not nearly expert enough to evaluate all the ramifications in all
the downstream CLDRs and systems. I'll be interested to see what
others discover.
Software should rely on tzData to determine an offset and when it
applies (or applied for historical time stamps). At this point I am
curious as to what will break the computers more, having a EST in tzData
that is not -05 or having tzData provide -05 or -04 instead of an alpha
formatted name abbreviation. Any time we have converted time from a
scientific standard such as milliseconds since an agreed on epoch to a
less precise form I'd expect (and accept) some errors when using less
precise notation.
I consider the "format" field to be a reference. In zone
America/New_York locals refer to their time as EST (or EDT 8 months a
year). "Eastern Time" year round. EDT is the standard time at this time
of year so I occasionally hear "EST" when people mean "EDT". If I looked
up America/Vancouver I would hope to see what locals called their time
(or what the National Research Council Canada calls the zone).
I don't consider nor expect "-04" to be common use for the eastern US or
"-07" to be common use for western Canada. If someone said it was 11:45
-07 instead of 11:45 PDT / 11:45 MST I would probably do bad math and
say "it is 4:45?". 14:45 -04? Anyone else thinking 10:45? So people
would need to add the UTC to the common use ... 11:45 UTC-07 or 14:45
UTC-04. That works but I don't see that becoming common.
Nearly all of my discussions with vendors and contacts referring to time
zones are based on reference cities. When we're setting a meeting we
will say New York time or Chicago time because those are more well known
than Indianapolis time or Dallas time.
I can see having %z or an explicit "-04" style format to see if that
breaks more systems than an EST that does not equal -05 but I hope that
the format field would change back to EST once software can handle the
legislated time zone (assuming EST is ever legislated as being -04).
[1] https://blog.poormansmath.net/the-time-it-takes-to-change-the-time/
includes a 2015 graphic showing offset from solar noon.
