DST (2018)

[vdeane]

Let me re-frame the problem:

Question: What time zone should a particular place observe?

Criteria:

1. No DST (the reason threads like this get started every spring and fall).

2. Time zone must be expressed as a whole-hour difference from UTC, for simplicity.

3. Within those two constraints, the indicated local time zone shall be selected as to minimize the sum of:

a. Total time before civil dawn occurring after 6:30am through the year
b. Total time after sunrise occurring before 5:30am through the year

4. The selected time zone may deviate from the indicated time zone to keep economically linked areas (CMSAs? TV markets?) on the same time zone.


The result will still probably be 4-5 time zones for the continental US...but the borders will be a little different.

I don't actually know what it would look like on a map.  That's part of why I'd love to take a weekend, play with the data, and see what happens.

BTW, before someone asks:

-- the use of dawn and sunrise in 3a/3b is intentional.  I'm trying to hypothesize criteria that allow for some ability to accommodate differences in latitude.

-- I'm not wedded to those specific times.  Frankly, I personally don't care since my daily schedule is linked more to sunrise than the clock, but I'm trying to recognize the two big complaints that always come up in time zone/DST discussions:  "I don't like kids going to school / getting ready for work in the dark" and "I'd rather have daylight after normal work hours than before".  I'm open to suggestions for alternatives.

This kind of exercise could also be used to describe when an optimal change between standard and daylight saving time ought to occur...but I won't go there since the time changes annoy me.

I wish I weren't running late on something for work.  This would be a kind of cool/fun programming/mapping weekend project.

I'm also curious to see how that would look worldwide.  Too bad I'm not quite sure of the easiest way to do it.  It sounds really interesting.

[MikeTheActuary]

Have I mentioned lately how much I hate Central time? 

But Central Time is God's Time!  Weekend afternoon sports naturally roll into primetime, and Carson's...er, Letterman's...er, Colbert's monologue wraps up at a reasonable bedtime, just as God intended. 

[kalvado]

And maintaining 1 hour granularity?
THat is what placing area into it straight time zone would do. No big optimization required.

Let me re-frame the problem:

Question: What time zone should a particular place observe?

Criteria:

1. No DST (the reason threads like this get started every spring and fall).

2. Time zone must be expressed as a whole-hour difference from UTC, for simplicity.

3. Within those two constraints, the indicated local time zone shall be selected as to minimize the sum of:

a. Total time before civil dawn occurring after 6:30am through the year
b. Total time after sunrise occurring before 5:30am through the year

4. The selected time zone may deviate from the indicated time zone to keep economically linked areas (CMSAs? TV markets?) on the same time zone.


The result will still probably be 4-5 time zones for the continental US...but the borders will be a little different.

I don't actually know what it would look like on a map.  That's part of why I'd love to take a weekend, play with the data, and see what happens.

BTW, before someone asks:

-- the use of dawn and sunrise in 3a/3b is intentional.  I'm trying to hypothesize criteria that allow for some ability to accommodate differences in latitude.

-- I'm not wedded to those specific times.  Frankly, I personally don't care since my daily schedule is linked more to sunrise than the clock, but I'm trying to recognize the two big complaints that always come up in time zone/DST discussions:  "I don't like kids going to school / getting ready for work in the dark" and "I'd rather have daylight after normal work hours than before".  I'm open to suggestions for alternatives.

This kind of exercise could also be used to describe when an optimal change between standard and daylight saving time ought to occur...but I won't go there since the time changes annoy me.

I wish I weren't running late on something for work.  This would be a kind of cool/fun programming/mapping weekend project.

Keep it simple. You basically require sunrise at equinox to be at 6 AM, that minimizes your unwanted hours.  That means putting noon as close to 11.59 AM as practical - that is, observing as-drawn time zone of closest (15 deg)*x meridian.
Some shift may occur since you define sunrise in 2 slightly different ways in (3)(a) and (3)(b), something like optimal noon at 12.15 PM -  but that would be totally absorbed by (4).

[MikeTheActuary]

Keep it simple. You basically require sunrise at equinox to be at 6 AM, that minimizes your unwanted hours.  That means putting noon as close to 11.59 AM as practical - that is, observing as-drawn time zone of closest (15 deg)*x meridian.
Some shift may occur since you define sunrise in 2 slightly different ways in (3)(a) and (3)(b), something like optimal noon at 12.15 PM -  but that would be totally absorbed by (4).

The difference in 3(a) and 3(b) should add a little bit of a longitudinal adjustment to it...but admittedly not a huge one at least as regards to CONUS.

I was also toying around with some alternative criteria...which would justify the fun of building a nifty macro-driven spreadsheet or bit of R code.  It wouldn't be as much fun if it were simplified. 

[tradephoric]

Suncalc.org is a great resource.  Click on any point in the world map and it will show sunrise, sunset, dusk, dawn, and sun duration times for that location.  It will also show what timezone the location is in and will tell you if it's on DST or not if you expand on "geodata for the selected location".  It also gives you sun paths and shadow lengths and solar eclipse info if you are interested.

https://www.suncalc.org/#/40.3549,-3.5815,5/2017.06.21/15:37/1/0

[kalvado]

Keep it simple. You basically require sunrise at equinox to be at 6 AM, that minimizes your unwanted hours.  That means putting noon as close to 11.59 AM as practical - that is, observing as-drawn time zone of closest (15 deg)*x meridian.
Some shift may occur since you define sunrise in 2 slightly different ways in (3)(a) and (3)(b), something like optimal noon at 12.15 PM -  but that would be totally absorbed by (4).

The difference in 3(a) and 3(b) should add a little bit of a longitudinal adjustment to it...but admittedly not a huge one at least as regards to CONUS.

I was also toying around with some alternative criteria...which would justify the fun of building a nifty macro-driven spreadsheet or bit of R code.  It wouldn't be as much fun if it were simplified. 

Well, here are some possibilities
(weight 10)  minimum of darkness after 8.30 AM
(weight 1) minimum of darkness after 7.30 or 8 AM
(weight  0.5) minimum of darkness after 6.30 AM
(weight 1) minimum of sunlight before 5 AM
(weight 2) minimum of darkness before 5 PM
(weight 0.2) minimum of darkness before 6.30 PM
(weight 0.2) minimum of sunlight after 10 PM


Alternatively, just use t^2 as a weight factor for unwanted lighting conditions. 

But everything will eventually boil down to "what is acceptable morning span"

[SP Cook]

MikeTheActuary expresses the formula better than anyone else has.   

As to CT vs ET or whatever.  I don't care that much.  I work a day job and that is not going to change.  My child is grown and does not catch a school bus.  I'm really not that much of an outdoor person.  The only real difference in time zones comes down to TV and TV sports.

If you live in ET or PT, "prime time" is 8-11 and Carson's pale imitators are 11:35.  If you live in CT or MT, then it is 7-10 and 10:35.  Both have their advantages.  There is something to being unrushed which ET gives you, but there is also something to being able to finish prime time and start the later shows w/o giving up 8 hours.

Sports are an issue.  It is great to get out west and see sports on weekdays in the early afternoon and on weekends to have games as early as 9 AM

[MikeTheActuary]

Well, here are some possibilities
(weight 10)  minimum of darkness after 8.30 AM
(weight 1) minimum of darkness after 7.30 or 8 AM
(weight  0.5) minimum of darkness after 6.30 AM
(weight 1) minimum of sunlight before 5 AM
(weight 2) minimum of darkness before 5 PM
(weight 0.2) minimum of darkness before 6.30 PM
(weight 0.2) minimum of sunlight after 10 PM


Alternatively, just use t^2 as a weight factor for unwanted lighting conditions. 

But everything will eventually boil down to "what is acceptable morning span"

True.

Or.... another set of critieria could be to maximize the number of hours that are "correct", say along these lines:

(using 24 hour times)

0030 = night
0130 = night
0230 = night
0330 = night
0430 = night
0530 = night or twilight
0630 = twilight or daylight
0730 = daylight
...
1630 = daylight
1730 = daylight or twilight
1830 = daylight or twilight
1930 = daylight or twilight
2030 = daylight or twilight or night
2130 = twilight or night
2230 = night
2330 = night

Perhaps give double- or triple-weight to the critical hours of morning commute and after-school+evening commute....

I think the end product with these criteria would end up looking like the 15° "ideal" zones shifted a bit to the west.

Oh yeah; this is SO going to the top of the "fun geek project when I have the time" pile. 

[kalvado]

I think the end product with these criteria would end up looking like the 15° "ideal" zones shifted a bit to the west.

Oh yeah; this is SO going to the top of the "fun geek project when I have the time" pile. 

Probably slightly bent bands compared to traditional north-south bands - to accommodate for more morning sunlight in  shorter winter days

[MikeTheActuary]

Thanks to some handy sunrise/sunset calculation functions I had tucked away, I now have a spreadsheet built that can be used to tackle the optimization I described in my last post.

I added weight to morning commute / afternoon / evening from what I detailed earlier.

I defined points and "ideal" lightness/darkness as follows:

0030, 0130, 0230, 0330, 0430 should be Night
0530 should be Night or Twilight
0600, 0630, 0645 should be Twilight or Day
0700, 0715, 0730, 0745, 0800, 0830 .... 1530, 1630 should be Day
1700, 1730, 1800, 1830, 1900, 1930, 20,00 should be Day or Twilight
2030 should be Day, Twilight, or Night
2130 should be Twilight or Night
2230, 2300 should be Night

I calculate the status of each of those points for each day of the year. I give one point for each point, each day.

My objective is to maximize the number of points a time zone selection yields.

Here's what I get for a handful of cities (with a caveat that the Excel VBA function I'm using for sunrise/sunset isn't perfect):

City: TimeZone [June sunrise/sunset] [December sunrise/sunset]

St. John's NF: UTC-3 [6/21:0433/2032] [12/21:0816/1642]
Bangor, Maine: UTC-4 [6/21:0449/2024] [12/21:0809/1657]
Washington, DC: UTC-5 [6/21:0443/1937] [12/21:0723/1649]
Toronto, ON: UTC-5 [6/21:0436/2003] [12/21:0748/1643]
Miami, FL: UTC-4 [6/21:0630/2015] [12/21:0803/1835]
Atlanta, GA: UTC-5 [6/21:0528/1952] [12/21:0739/1733]
Indianapolis, IN: UTC-5 [6/21:0517/2016] [12/21:0802/1724]
Thunder Bay, ON: UTC-5 [6/21:0456/2102] [12/21:0846/1705]
Memphis, TN: UTC-6 [6/21:0446/1918] [12/21:0705/1652]
New Orleans, LA: UTC-6 [6/21:0500/1904] [12/21:0652/1705]
Dallas, TX: UTC-6 [6/21:0520/1939] [12/21:0726/1725]
Odessa, TX: UTC-6 [6/21:0545/1958] [12/21:0746/1750]
Denver, CO: UTC-6 [6/21:0532/2031] [12/21:0817/1739]
Saskatoon, SK: UTC-6 [6/21:0446/2132] [12/21:0913/1656]
Salt Lake City, UT: UTC-7 [6/21:0456/2002] [12/21:0749/1703]
Phoenix, AZ: UTC-7 [6/21:0519/1942] [12/21:0729/1725]
Edmonton, AB: UTC-7 [6/21:0404/2107] [12/21:0848/1616]
Los Angeles, CA: UTC-7 [6/21:0542/2008] [12/21:0755/1748]
Seattle, WA: UTC-7 [6/21:0512/2111] [12/21:0855/1721]
San Francisco, CA: UTC-8 [6/21:0448/1935] [12/21:0722/1654]
Anchorage, AK: UTC-9 [6/21:0320/2243] [12/21:1014/1542]
Honolulu HI: UTC-10 [6/21:0551/1916] [12/21:0705/1755]

I think I might need to give a bit more weight to the morning criteria.

[20160805]

I know this will probably get buried, but as someone who goes to bed pretty early (21:00 or just after), I see no point in gratuitous evening daylight and thus favour standard time, which for my location is GMT-6.  That would make our June daylight roughly 4:00-19:40 and December 7:25-16:15, and it would eliminate late-evening sleep problems, problems associated with changing the clock, and the oddity of the latest sunrise being in early November.

[MikeTheActuary]

I played with my algorithm a bit more, aiming for results that, for places on the whole multiples of 15° longitude, the result (before rounding) would be a half-hour later than standard time (a logical compromise for ST and DST), and that at high latitudes things wouldn't get too screwy.

New algorithm:

For each day of the year, for each of the following time points, award the following points:
0030, 0130, 0230, 0330, 0430:  10 points if night
0530: 4 points if night, 4 points if twilight, 2 points if day
0600: 3-1/3 points regardless
0630: 2 points if night, 4 points if twilight, 4 points if day
0645, 0700: 5 points if twilight, 5 points if day
0715: 4 points if twilight, 6 points if day
0730: 2 points if twilight, 8 points if day
0745, 0800, 0830 [and every xx30 until] 1630, 1700: 10 points if day
1730: 1 point if twilight, 9 points if day
1800: 3 points if twilight, 7 points if day
1830: 4 points if twilight, 6 points if day
1900: 1 point if night, 4 points if twilight, 5 points if day
1930: 2 points if night, 4 points if twilight, 4 points if day
2000: 3-1/3 points
2030: 4 points if night, 4 points if twilight, 2 points if day
2130, 2230, 2330: 10 points if night

...which seems to do a pretty good job of balancing "commute in some light", "'extra' daylight in evening in the summer" (sorry former RandomDude!), and "no time change".

If some of the weightings seem a little odd...well, there was trial and error at work. 

I ran it through the same selection of cities, and quickly observed something: if you don't want an indicated time-zone boundary running through the Northeast Corridor, the "whole hour offsets from UTC" has to be relaxed.

So, here's what my spreadsheet is putting out now (again with the caveat that the sunrise/sunset calculations are imperfect):

St. John's NF: UTC-3.5 [6/21:0403/2002] [12/21:0746/1612]   (Current = [6/21: 0503/2102] [12/21: 0746/1612])
Bangor, Maine: UTC-4.5 [6/21:0419/1955] [12/21:0740/1627]   (Current = [6/21: 0449/2025] [12/21: 0710/1557])
Washington, DC: UTC-4.5 [6/21:0513/2007] [12/21:0753/1720]   (Current = [6/21: 0543/2037] [12/21: 0723/1649])
Toronto, ON: UTC-5.5 [6/21:0406/1933] [12/21:0718/1613]   (Current = [6/21: 0536/2103] [12/21: 0748/1643])
Miami, FL: UTC-4.5 [6/21:0600/1945] [12/21:0733/1805]   (Current = [6/21: 0630/2015] [12/21: 0703/1735])
Atlanta, GA: UTC-5.5 [6/21:0457/1922] [12/21:0709/1703]   (Current = [6/21: 0628/2052] [12/21: 0739/1733])
Indianapolis, IN: UTC-5.5 [6/21:0447/1946] [12/21:0733/1653]   (Current = [6/21: 0617/2116] [12/21: 0802/1724])
Thunder Bay, ON: UTC-5.5 [6/21:0425/2032] [12/21:0816/1634]   (Current = [6/21: 0456/2102] [12/21: 0746/1605])
Memphis, TN: UTC-5.5 [6/21:0516/1948] [12/21:0735/1722]   (Current = [6/21: 0546/2018] [12/21: 0705/1652])
New Orleans, LA: UTC-5.5 [6/21:0530/1935] [12/21:0722/1735]   (Current = [6/21: 0600/2004] [12/21: 0652/1705])
Dallas, TX: UTC-5.5 [6/21:0550/2008] [12/21:0756/1755]   (Current = [6/21: 0620/2039] [12/21: 0726/1725])
Odessa, TX: UTC-6.5 [6/21:0515/1929] [12/21:0716/1720]   (Current = [6/21: 0645/2058] [12/21: 0746/1750])
Denver, CO: UTC-6.5 [6/21:0502/2001] [12/21:0748/1709]   (Current = [6/21: 0532/2032] [12/21: 0718/1639])
Saskatoon, SK: UTC-6.5 [6/21:0415/2101] [12/21:0843/1626]   (Current = [6/21: 0446/2132] [12/21: 0913/1656])
Salt Lake City, UT: UTC-6.5 [6/21:0527/2032] [12/21:0818/1734]   (Current = [6/21: 0556/2102] [12/21: 0749/1703])
Phoenix, AZ: UTC-6.5 [6/21:0549/2011] [12/21:0758/1755]   (Current = [6/21: 0519/1942] [12/21: 0729/1725])
Edmonton, AB: UTC-6.5 [6/21:0434/2138] [12/21:0919/1646]   (Current = [6/21: 0504/2207] [12/21: 0848/1616])
Los Angeles, CA: UTC-7.5 [6/21:0512/1938] [12/21:0725/1718]   (Current = [6/21: 0542/2008] [12/21: 0655/1648])
Seattle, WA: UTC-7.5 [6/21:0441/2041] [12/21:0825/1650]   (Current = [6/21: 0511/2111] [12/21: 0755/1620])
San Francisco, CA: UTC-7.5 [6/21:0518/2005] [12/21:0752/1725]   (Current = [6/21: 0548/2035] [12/21: 0722/1654])
Anchorage, AK: UTC-8.5 [6/21:0350/2313] [12/21:1044/1611]   (Current = [6/21: 0420/2343] [12/21: 1014/1542])
Honolulu HI: UTC-9.5 [6/21:0620/1947] [12/21:0735/1825]   (Current = [6/21: 0551/1916] [12/21: 0705/1755])

[mrsman]

Thanks to some handy sunrise/sunset calculation functions I had tucked away, I now have a spreadsheet built that can be used to tackle the optimization I described in my last post.

I added weight to morning commute / afternoon / evening from what I detailed earlier.

I defined points and "ideal" lightness/darkness as follows:

0030, 0130, 0230, 0330, 0430 should be Night
0530 should be Night or Twilight
0600, 0630, 0645 should be Twilight or Day
0700, 0715, 0730, 0745, 0800, 0830 .... 1530, 1630 should be Day
1700, 1730, 1800, 1830, 1900, 1930, 20,00 should be Day or Twilight
2030 should be Day, Twilight, or Night
2130 should be Twilight or Night
2230, 2300 should be Night

I calculate the status of each of those points for each day of the year. I give one point for each point, each day.

My objective is to maximize the number of points a time zone selection yields.

Here's what I get for a handful of cities (with a caveat that the Excel VBA function I'm using for sunrise/sunset isn't perfect):

City: TimeZone [June sunrise/sunset] [December sunrise/sunset]

St. John's NF: UTC-3 [6/21:0433/2032] [12/21:0816/1642]
Bangor, Maine: UTC-4 [6/21:0449/2024] [12/21:0809/1657]
Washington, DC: UTC-5 [6/21:0443/1937] [12/21:0723/1649]
Toronto, ON: UTC-5 [6/21:0436/2003] [12/21:0748/1643]
Miami, FL: UTC-4 [6/21:0630/2015] [12/21:0803/1835]
Atlanta, GA: UTC-5 [6/21:0528/1952] [12/21:0739/1733]
Indianapolis, IN: UTC-5 [6/21:0517/2016] [12/21:0802/1724]
Thunder Bay, ON: UTC-5 [6/21:0456/2102] [12/21:0846/1705]
Memphis, TN: UTC-6 [6/21:0446/1918] [12/21:0705/1652]
New Orleans, LA: UTC-6 [6/21:0500/1904] [12/21:0652/1705]
Dallas, TX: UTC-6 [6/21:0520/1939] [12/21:0726/1725]
Odessa, TX: UTC-6 [6/21:0545/1958] [12/21:0746/1750]
Denver, CO: UTC-6 [6/21:0532/2031] [12/21:0817/1739]
Saskatoon, SK: UTC-6 [6/21:0446/2132] [12/21:0913/1656]
Salt Lake City, UT: UTC-7 [6/21:0456/2002] [12/21:0749/1703]
Phoenix, AZ: UTC-7 [6/21:0519/1942] [12/21:0729/1725]
Edmonton, AB: UTC-7 [6/21:0404/2107] [12/21:0848/1616]
Los Angeles, CA: UTC-7 [6/21:0542/2008] [12/21:0755/1748]
Seattle, WA: UTC-7 [6/21:0512/2111] [12/21:0855/1721]
San Francisco, CA: UTC-8 [6/21:0448/1935] [12/21:0722/1654]
Anchorage, AK: UTC-9 [6/21:0320/2243] [12/21:1014/1542]
Honolulu HI: UTC-10 [6/21:0551/1916] [12/21:0705/1755]

I think I might need to give a bit more weight to the morning criteria.

This is amazing work.  Thank you.

If I had to make adjustments to this, I would put in a grand rule of whole number hours from UTC only
and winter solstice sunrise between 0700 and 0759 [excluding Alaska], which would do the following to your list:

Cities listed in absolute order of sunrise, from UTC, on the winter solstice:

ATL STD TIME
St. John's NF: UTC-4 [6/21:0333/1932] [12/21:0716/1542]

EST
Miami, FL: UTC-5 [6/21:0530/1915] [12/21:0703/1735]
Bangor, Maine: UTC-5 [6/21:0349/1924] [12/21:0709/1557]
Washington, DC: UTC-5 [6/21:0443/1937] [12/21:0723/1649]
Atlanta, GA: UTC-5 [6/21:0528/1952] [12/21:0739/1733]
Toronto, ON: UTC-5 [6/21:0436/2003] [12/21:0748/1643]
New Orleans, LA: UTC-5 [6/21:0600/2004] [12/21:0752/1805]

CST
Indianapolis, IN: UTC-6 [6/21:0417/1916] [12/21:0702/1624]
Memphis, TN: UTC-6 [6/21:0446/1918] [12/21:0705/1652]
Dallas, TX: UTC-6 [6/21:0520/1939] [12/21:0726/1725]
Thunder Bay, ON: UTC-6 [6/21:0356/2002] [12/21:0746/1605]
Odessa, TX: UTC-6 [6/21:0545/1958] [12/21:0746/1750]

MST
Denver, CO: UTC-7 [6/21:0432/1931] [12/21:0717/1639]
Phoenix, AZ: UTC-7 [6/21:0519/1942] [12/21:0729/1725]
Salt Lake City, UT: UTC-7 [6/21:0456/2002] [12/21:0749/1703]
Los Angeles, CA: UTC-7 [6/21:0542/2008] [12/21:0755/1748]

PST
Saskatoon, SK: UTC-8 [6/21:0246/1932] [12/21:0713/1456]
San Francisco, CA: UTC-8 [6/21:0448/1935] [12/21:0722/1654]
Edmonton, AB: UTC-8 [6/21:0304/2007] [12/21:0748/1516]
Seattle, WA: UTC-8 [6/21:0412/2011] [12/21:0755/1621]


HAWAII TIME
Honolulu HI: UTC-10 [6/21:0551/1916] [12/21:0705/1755]

ALASKA TIME
Anchorage, AK: UTC-9 [6/21:0320/2243] [12/21:1014/1542]

(Hopefully my adjustments did not mess up your calculaions Mike, please check my work.)

[tradephoric]

PST
Saskatoon, SK: UTC-8 [6/21:0246/1932] [12/21:0713/1456]
San Francisco, CA: UTC-8 [6/21:0448/1935] [12/21:0722/1654]
Edmonton, AB: UTC-8 [6/21:0304/2007] [12/21:0748/1516]
Seattle, WA: UTC-8 [6/21:0412/2011] [12/21:0755/1621]

Saskatoon looks a little out of wack.  Currently in Saskatoon dawn begins at 03:56 and dusk ends at 22:20 during the summer solstice.  Under your proposal, dawn begins at 01:56 and dusk ends at 20:20.  Saskatoon being that far north, dusk shouldn't be ending at 8:20PM during the summer.

[kalvado]

PST
Saskatoon, SK: UTC-8 [6/21:0246/1932] [12/21:0713/1456]
San Francisco, CA: UTC-8 [6/21:0448/1935] [12/21:0722/1654]
Edmonton, AB: UTC-8 [6/21:0304/2007] [12/21:0748/1516]
Seattle, WA: UTC-8 [6/21:0412/2011] [12/21:0755/1621]

Saskatoon looks a little out of wack.  Currently in Saskatoon dawn begins at 03:56 and dusk ends at 22:20 during the summer solstice.  Under your proposal, dawn begins at 01:56 and dusk ends at 20:20.  Saskatoon being that far north, dusk shouldn't be ending at 8:20PM during the summer.

Effect of hard anchoring latest sunrise time. Shift of sunrise time is most noticeable in areas further north.  Ignoring civil dusk part and some other fine print - if you have X ours of daylight in winter and X hours of night in summer, and hard-anchor sunrise to a certain time T, shift of dawn time is (12-X) hours. X is 8.5-9 hours i in northern part of US, and 3-3.5 shift  is almost tolerable. further north it breaks down.

[tradephoric]

Lawmakers in South Carolina are trying to get an advisory referendum on the November ballot to ask voters if they want to end spring forward and fall back and observe year-round daylight saving time. 

https://www.usnews.com/news/best-states/south-carolina/articles/2018-04-25/voters-may-decide-future-of-daylight-saving-time-in-november

Seems like more and more state legislators are favoring the idea of year-round daylight savings time.  This is a positive development, as getting rid of daylight saving time would completely limit state choice on this issue.  Under the current system, all 50 states have the choice to either observe DST or stay on standard time year-round.  Given that choice, 2 states have decided to remain on standard time year round (Arizona and Hawaii).  So why should we force the 48 states that have decided to be on DST to go back to standard time?  Also, it would be a mistake to force Arizona and Hawaii to go to year-round DST.  Even though they are well in the minority in not observing DST, they should have the right to remain on standard time if they feel like it's the best thing for their state.  So if year-round DST passes and a state senator is up in arms about the kids getting on the bus in the dark, then that individual state can debate whether they want to stay on year-round DST or go to year-round standard time.

[kalvado]

Lawmakers in South Carolina are trying to get an advisory referendum on the November ballot to ask voters if they want to end spring forward and fall back and observe year-round daylight saving time. 

https://www.usnews.com/news/best-states/south-carolina/articles/2018-04-25/voters-may-decide-future-of-daylight-saving-time-in-november

Seems like more and more state legislators are favoring the idea of year-round daylight savings time.  This is a positive development, as getting rid of daylight saving time would completely limit state choice on this issue.  Under the current system, all 50 states have the choice to either observe DST or stay on standard time year-round.  Given that choice, 2 states have decided to remain on standard time year round (Arizona and Hawaii).  So why should we force the 48 states that have decided to be on DST to go back to standard time?  Also, it would be a mistake to force Arizona and Hawaii to go to year-round DST.  Even though they are well in the minority in not observing DST, they should have the right to remain on standard time if they feel like it's the best thing for their state.  So if year-round DST passes and a state senator is up in arms about the kids getting on the bus in the dark, then that individual state can debate whether they want to stay on year-round DST or go to year-round standard time.

Making it short: currently, federal agency has the power to determine time zone for the state (or part thereof). Look like states want some of that power in their hands.

[tradephoric]

^The federal government simply sets the duration of DST under the Uniform Time Act of 1966.  The act was last amended in 2005 when Busch increased Daylight Saving Time by 4 weeks.  Ultimately, amending the Uniform Time Act of 1966 to make DST year-round wouldn't be giving more power to the states (they would still decide if they observe DST or not), but abolishing DST entirely would most certainly be taking power away from the states (states would be forced into standard time... like it or not).     

[1995hoo]

Kind of hard to believe that prior to 1987, the clocks would have just gone ahead yesterday (last Sunday in April). I think the current US DST schedule makes the change too soon, but the end of April was quite late.

[US 89]

The easiest way to observe permanent DST is to move one time zone east and not observe DST in that zone. Saskatchewan already does this; it is essentially in the Mountain zone with permanent DST. By being on CST year round, they are basically on permanent MDT. Depending on how you look at it, Arizona does the same thing. By observing MST year round, they are essentially on PDT.

Kind of hard to believe that prior to 1987, the clocks would have just gone ahead yesterday (last Sunday in April). I think the current US DST schedule makes the change too soon, but the end of April was quite late.

Agreed. IMO, DST should begin the first Sunday of April and end the second-to-last Sunday of October.

[mrsman]

PST
Saskatoon, SK: UTC-8 [6/21:0246/1932] [12/21:0713/1456]
San Francisco, CA: UTC-8 [6/21:0448/1935] [12/21:0722/1654]
Edmonton, AB: UTC-8 [6/21:0304/2007] [12/21:0748/1516]
Seattle, WA: UTC-8 [6/21:0412/2011] [12/21:0755/1621]

Saskatoon looks a little out of wack.  Currently in Saskatoon dawn begins at 03:56 and dusk ends at 22:20 during the summer solstice.  Under your proposal, dawn begins at 01:56 and dusk ends at 20:20.  Saskatoon being that far north, dusk shouldn't be ending at 8:20PM during the summer.

Effect of hard anchoring latest sunrise time. Shift of sunrise time is most noticeable in areas further north.  Ignoring civil dusk part and some other fine print - if you have X ours of daylight in winter and X hours of night in summer, and hard-anchor sunrise to a certain time T, shift of dawn time is (12-X) hours. X is 8.5-9 hours i in northern part of US, and 3-3.5 shift  is almost tolerable. further north it breaks down.

The point of the exercise is to pick one time constraint.  In my example, I forced sunrise on the winter solstice to always be between 7:00 and 7:59.  But the exercise can also be done with respect to sunset or high noon as well or by focusing on the effect at the summer solstice or the equinoxes.  Since most of the DST complaints tend to involve dark mornings, I figured the most important factor for people was being sure that school kids weren't in the dark so the sun has to rise before 8:00.  But other factors may take precedence.

IMO pick one time and stick with it.  School can always start later to avoid the dark mornings.

It's also interesting to see Miami:

Miami, FL: UTC-5 [6/21:0530/1915] [12/21:0703/1735]
if it went on permanent DST would look like:

Miami, FL: UTC-4 [6/21:0630/2015] [12/21:0803/1835] which actually does extend sunshine during the winter season till almost 7:00 PM.  The problem is that the western part of the state will have a very late sunrise duing the winter, close to 9:00 am.

[tradephoric]

Assume an early riser is awake from 5AM-9PM and a late riser is from 8AM-midnight.  Under standard time during the winter solstice, a late riser in Boston experiences 434 minutes of darkness during their waking hours compared to just 352 minutes of darkness for the early riser.  Now if permanent DST was enacted, a late riser in Boston would only experience 374 minutes of darkness during their waking hours and the early riser would still experience the same 352 minutes of darkness.  Permanent DST would minimize the amount of darkness people receive during their waking hours in the winter.  As you can see, even with permanent DST, late risers in Boston miss out on 22 minutes of light vs. their earlier riser counterparts, but that's a lot better than missing out on 82 minutes of light.  Obviously, the farther west you go, the later the sunrise gets which helps ensure that it's still dark when a late riser awakens (which is a good thing, because there is so precious little sun in the winter that you don't want to waste it sleeping!).  By the time you get to Syracuse, both late risers and early risers experience the same minutes of darkness during their waking hours during the winter. 

[webny99]

Another foreseeable issue is that some time zones (including our Eastern Time Zone) are wider than 1/24 of the earth. That is to say, the extremities cannot both have a sunrise between 7:00 and 7:59. I can see this being an issue with Indiana and Maine.

[hotdogPi]

Another foreseeable issue is that some time zones (including our Eastern Time Zone) are wider than 1/24 of the earth. That is to say, the extremities cannot both have a sunrise between 7:00 and 7:59. I can see this being an issue with Indiana and Maine.

Indiana can go to Central. It would split Kentucky in half, and hbelkins would still be in Eastern.

[kalvado]

Another foreseeable issue is that some time zones (including our Eastern Time Zone) are wider than 1/24 of the earth. That is to say, the extremities cannot both have a sunrise between 7:00 and 7:59. I can see this being an issue with Indiana and Maine.

More like an issue with people having to suck it up and stop bitching about dawn being 10 minutes later than they would love to.
One  hour granularity of clock settings allows only that much of tweaking. Complains about day being too short in winter may be well founded on exact data - but tweaking clock is not going to resolve it.
Just FYI: Hawaii have no such issues, neither does southern Mexico..