[repsilat]

One real problem with daylight savings is that it goes the wrong way. The idea was that in the winter there would be fewer hours with the sun up, and it made sense to put those hours in a more useful part of the day.

When daylight savings was introduced, that "more useful part of the day" was earlier in the day. That doesn't hold today, though - as TFA says, most people work indoors under artificial lighting, and shifting daylight hours earlier into the day just means they're wasted while we're behind desks. What we really need is more daylight in the evening, so we can make productive use of our leisure time playing catch with our kids, practicing with the football club, drinking beer on the porch etc.

Still, killing daylight savings would probably be better than "reverse daylight savings", and more politically feasible. Certainly more feasible than crazy ideas like moving to UTC (and probably better for the average person anyway).

My pet peeve, though, is leap-seconds. Life would be a lot easier if you could rely on days always having 86400 seconds.

[haberman]

> The idea was that in the winter there would be fewer hours with the sun up, and it made sense to put those hours in a more useful part of the day.

Daylight Savings time happens during the summer, not the winter. Winter time is unadjusted and tracks a true solar day. Summer time is adjusted, and indeed it's adjusted in the direction you favor (to provide more daylight hours at night).

[seanmcdirmid]

I live in China and we don't have daylight savings time. The sun rises at around 4AM in the summer, which really messes up my sleep habits if I forget to close the blinds before bed. Winters are about right.

[wr1472]

Your problem is much more easily solved: remember to close the blinds.

[mistercow]

>Winter time is unadjusted and tracks a true solar day.

That really depends on where you are in the time zone. If you're right at the edge, you may be closer to tracking a true solar day in the summer.

[Firehed]

Tracks a true solar day? Is there actually a scientific definition of when a day starts and ends? As far as I can tell skimming Wikipedia, it's simply 86400 seconds.

Certainly by my body clock, a "true solar day" is when I'm on DST.

[mooism2]

Midday GMT is defined to be when the mean sun is directly over the prime meridian.

("Mean sun" as opposed to "apparent sun" because of the Earth being a naturally occurring celestial object, rather than an idealised mathematical model. See http://en.wikipedia.org/wiki/Equation_of_time .)

[jey]

> My pet peeve, though, is leap-seconds. Life would be a lot easier if you could rely on days always having 86400 seconds.

Is this sarcasm? The only time I've ever noticed leap seconds was the story about a bug in the Linux kernel's handling of them.

[anonymouz]

There is a good reason for leap seconds though: They make sure that UTC (based on atomic clocks) stays in sync with the solar day. If we didn't have those, the two would slowly drift apart and we would have to introduce a bigger jump at some later point.

[perlgeek]

By why is it important that UTC stays in sync with the solar day? The drift is very small, and I see no practical application for a sync in such small amounts.

Most people's local time isn't synced with the solar time anyway (that only happens in the middle of a time zone, and only if there's no crazy stuff like DST going on).

The only people who care about that are probably the astronomy guys, and they already need to adjust their time with other stuff for most observations.

[Benferhat]

> There is a good reason for leap seconds though

Isn't that the only reason for them?

[gpvos]

> My pet peeve, though, is leap-seconds. Life would be a lot easier if you could rely on days always having 86400 seconds.

Is that truly a problem in your daily life, or just when you deal with computers? In the latter case, just make those computers follow TAI instead of UTC.

For the foreseeable future, I'm certain that humans will prefer the time that the sun stands highest in the sky to be called 12:00. Even though it's pretty far in the future when it will be noticeable, I think it's pretty presumptuous to burden our progeny with something else.

[zanny]

There isn't anything saying a business can't have employees work 8pm - 4am during the winter. Besides the fact a lot of people can't adapt their sleep schedule.

[npsimons]

Try telling the businesses that. Far too many businesses (especially entrenched, large employers) are stuck in the 8 to 5, 5 days a week schedule. Allowing more flexible schedules would be a boon all around (less traffic congestion, allow employees to work when maximally effective, etc). But even places that pay lip service to flex time are ironically inflexible; usually they just have slightly different hours, but everyone is expected to stick to those hours, or they lose face (bye-bye promotions).

[codex]

DST doesn't shift daylight hours earlier in the day; rather, it ensures that the sun rises at approximately the same time every morning, regardless of season.

[Someone]

Time of sunrise various much more than an hour once you are far enough away from the equator. For example, http://aa.usno.navy.mil/data/docs/RS_OneYear.php gives over 2 hours for Texas, over 3 hours for New York and almost 4 hours for Washington.

[dbaupp]

And 24 hours once you are above/below the Arctic/Antarctic circles.

[codex]

Yes, hence the "approximately." The current system is a compromise; you can't go slewing the time continuously across latitudes and day of year, at least not yet, but you can't shock people with three hour jumps, either. I for one would welcome two dimensional time zones and continuous slewing. In an era of smart timekeeping, anything is possible.

[nitrogen]

I would much rather have sunset, rather than sunrise, synchronized. If the sun always set at 10PM, life would be easy.

[kooshball]

I think what the OP is looking for, is for DST to ensure that the sun sets at the same time every evening.

[rms]

Yeah, I would have a strong preference for permanent reverse daylight savings.