Continuous Daylight Saving Time

Daylight Saving Time seems to serve three major purposes:

  • Health: keeping sunrise roughly constant relative to when work or school starts makes modern routines easier on our circadian rhythms, improving our pyschological health and perhaps also our physical health. In addition, the daylight "saved" by not "sleeping in" hours past sunrise during the summer makes more outdoor activity possible, increasing the amount of exercise we get without conscious effort.
  • Energy use: By using less artificial light and spending less time inside watching TV during the summer, America saves about 1% on total energy use by using Daylight Saving Time.
  • Safety: Daylight Saving Time tries to keep both morning and evening commutes in daylight when possible. But when that isn't possible, it tries to ensure that at least the morning commute is during daylight. This reduces car-accident injuries by thousands or tens of thousands per year.

I think a time system could improve health, energy use, and safety even more if it were to make small adjustments throughout the year instead of large adjustments twice a year. For example, a small amount of time might be added or taken away just before 2am every morning, in order to keep sunrises at 6am at a latitude of 40 degrees. The daily changes would be small enough for most people to ignore -- less than two minutes per day even around the equinoxes.

Interestingly, switching to continuous time change would also address the main criticisms of DST:

  • Lost productivity and an increase in fatal auto accidents twice a year due to disruption of sleeping patterns.
  • Lost productivity fiddling with clocks.
  • Farmers are forced out of synchronization with the rest of society.

It seems like my favorite kind of compromise, one that reveals a false trade-off and makes both sides happier than they would have been with their previous preferred solutions.

Of course, there would be new drawbacks. Certain time calculations would be more difficult: night-shift workers might find themselves needing to keep track of the changing length of each day, instead of being confused only twice a year. Planning a weekly meeting involving people in different hemispheres (or DST regimes) would become more difficult, especially if people on each hemisphere have tight schedules.

We would also have to replace our clocks and watches. I'm not about to pretend that forcing everyone to purchase new clocks would be a good thing by itself, but at least it would only be a one-time cost; computing power is cheap enough that the the price of clocks would not increase permanently. When we upgrade our clocks to deal with days that vary slightly in length, we should also give them all the ability to update themselves; this would be more pleasant than requiring you to enter the date in addition to the time after each power outage. We could also dramatically improve the user interfaces of most alarm clocks with respect to how often they fail to wake people up, but that's the subject for another blog post.

This "Continuous DST" proposal is not to be confused with the proposal known as "Year-round DST". The advantages of DST arise from the twice-yearly changes to our clocks corresponding to the changes in the seasons. While "year-round DST" might make sense as a short-term response to an energy crisis such as World War II, in the long term it equivalent to not having DST at all: over a period of several years, everyone will shift their hours back to when they are comfortable being awake unless the government also legislates working hours, store hours, and prime-time television.

I'll admit to being atypical when it comes to sleeping schedules. I work from home and can keep almost any schedule I want. I tend to be most productive at nights, when there are few distractions, so I often sleep during the day. I prefer to be outside during the evening and night, when I don't have to wear sunglasses. (As an added bonus, when I go grocery shopping, my dairy products will take less damage from the walk home). On the other hand, in college, when many students wouldn't even consider taking a class before 10am, I didn't mind having an 8am MWF class as long as I also had a 8:10am class on Tuesday and Thursday.

I'm sure many readers do keep "normal hours", whether by coercion or choice, so what do you think of Continuous DST?

37 Responses to “Continuous Daylight Saving Time”

  1. Max Says:

    The lost productivity due to fiddling with clocks doesn’t change any with your system; we could all get fancy new self setting clocks without changing DST. The idea that farmers are harmed because the rest of society is ‘only’ available to them during 7(10? 12?) hours that vary during the year strikes me as silly, but perhaps I am missing some important ramification of the time change.

    The car accident thing is interesting, but I have had conversations with (several!) people who fell asleep and drove off the road, mostly swing shift factory people, but creating self assessment tools and encouraging safer habits in general would probably have the same impact as the mild shift.

    I’m not sure if it would be irritating or not, but the current imbalance between DST and normal months would mean that the length of the day would change more during the winter. Only a little, but if it was noticeable…

    I’m a fan of just going to DST full year, mostly because during the summer, I would rather work in the dark than play in the dark and during winter it hardly matters. I get the busing issue, but if we listen to psychologists, children probably shouldn’t bet getting to school much before 10:00 AM anyway, too bad that would inconvenience so many parents.

    The benefits of cooperating on business hours are getting smaller, especially as the hours that businesses stay open increases. To the extent that it is mostly just a giant mind hack “You will pretend it is earlier/later.” I’m not convinced it is really all that necessary.

  2. Sebastian Says:

    Why not drop that time-shift-shit completely? It worked perfectly without for decades.

  3. Johan Sundström Says:

    How would continuous DST to work, on a global perspective? (As in “in San Francisco, in New York and in Beijing”, for instance; the three cities are unlikely to want to share time readings, given their different angles to the sun.)

    Keep current time zones intact all over the world, reset to a common DST/non-DST base reading and start adjusting towards the sliding time at 02, local time? To heck with time zones, define the pan-earth time function f(atomic clock UT1 time, lat, lng) = t?

    I think introducing Stardates, and facing the societal impacts and adjustments that would have, would be a more sane course of action, and then I’m not much of a trekkie, either. I find this proposition yet another layer of complexity and software bugs (Y2K is nothing in comparison) on top of an already broken system, for very little gain and much additional pain.

  4. John Says:

    I’ve actually always been a proponent of abolishing DST completely, mostly because I don’t like having to remember whether it’s on or not to know whether I’m GST-6 or GST-5. I believe worldwide collaboration would be quite a bit simpler if local time systems didn’t keep jumping around inconsistently. DST just seems somehow inelegant.

    That said, if you think the benefits of it are worth the complication, I’d tend to agree that a more finely graded system would have certain advantages over the current one. OTOH, I don’t think that a continuous system is the best answer, mostly due to the clock problem you point out.

    I would recommend a finer-grained system, maybe consisting of 4 30-minute adjustments (one every 13 weeks) or 12 10-minute adjustments (every month?), except that the complexity of the system would increase much further the difficulty of reconciling different time convesions than the current (mostly) 1-hour off systems in place.

    After reflecting on it a while, I’ve decided that I think the current system is close to the best possible pragmatic approach to daylight saving time. The calculation is easy for almost anyone to understand (if not remember…), and the majority of potential benefits to DST are realized.

    At some point, systems just get so entrenched that altering them, even for the better, would create more problems than it solves. It seems to me that the best thing to do, when that point is reached, is offer an alternative rather than an incremental improvement, but I have a hard time seeing any alternative time system ever being accepted either.

  5. Drew Thaler Says:

    It’s an interesting idea. Horribly impractical, though. :-)

    In a nutshell, the problem is that we seem to want an absolute, fixed scale (9:00am) to refer to approximately the same time of day all year long … where “time of day” is a relative, variable value. For example: 9am is about two or three hours after sunrise at our latitude. When you state the problem that way, you can see a couple of different classes of solutions:

    (1) change our measurements: have work start at 8am sometimes and 9am sometimes. This might be the best answer, but nobody really seems to want this one.

    (2) change our rulers: redefine 9am so that it’s what we expect it to be. DST in all its forms is an example of this.

    (3) use a more appropriate measure: come up with a sunrise-relative time specification and use that instead. Instead of work starting at ’9:00am’, it might instead start at ’2:00pd’ (post diliculum).

    I think #3 is interesting too. And it’s the kind of thing that tends to be done in computer science. If one measure turns out to be inappropriate, well, come up with another one that fits better!

    Of course, a time method which is non-absolute like that is just a general problem. Businesses which are further north than others would open at different times, and you don’t really want the time differences to be fully analog… so you’d probably need to introduce vertical time zones as well as horizontal ones. And many things would still need an absolute time scale, so you’d never be able to ditch the old am/pm scheme.

    In the end, though, it’s probably too many complications. I have to agree with John — as much as the current DST system sucks it’s really not a bad compromise.