For a couple years now, this has been a term in my friends' and my lexicon. It came up first at one of our late night Denny's runs. We never went before 10, and typical was 11-1, which meant getting back even later. On the way home, as we were talking about the next day, someone pointed out that, being after midnight, that it was actually today. This is a prime example of today-tomorrow unreasonable. This is really a precursor to the discussion of when one day ends and the next starts.

The concept of the day and the year, both based on astronomical events, makes perfect sense. Hours, however, I feel are a big more arbitrary. We are diurnal creatures and therefore daylight definately has define the way we live before and after the invent of accessable artificial light. The current method of keeping time is designed to center around noon, the center of the daylight hours. This doesn't seem most appropriate to me, as we are much more tuned to sunrise than sunset. People generally wake up around sunrise, or at least the morning is the time around when the sun rises. Most people stay awake much longer than daylight which means that sunset has very little effect on the human cycle.

I believe there is a natural human daily cycle, also based on the sun as the clock is, but with a different concept of when the beginning of the day is than the clocks notion of midnight. In general most people talk about "the beginning of their day" as when they wake up. No one I know who goes to bed at 1am and wakes up at 9am talkes about the previous midnight when they were still awake from the day before as the beginning of their day. I believe people who at 1am refer to what they will due after they wake up (after they go to bed) as today as opposed to tomorrow are today-tomorrow unreasonable. When the clock rolls over from 11:59pm to 12am, what you did 5 minutes ago doesn't suddenly become what you did, yesterday night; It was tonight. Never-the-less, this can get tricky. When you schedule to do something at 1am Saturday, do you mean actually Saturday (as in an extention of Friday night but after midnight) or do you mean Saturday night after midnight (techincally Sunday)? I prefer to think of things in terms of 1am Saturday night meaning the night after Saturday, but you never know how other people might take it.

Furthermore, we generally consider night to follow day. This is probably because when we wake it is day and then at some point while awake it becomes night (still part of the same day). At some point you go to sleep and when you wake up it is daytime again, and the next calendar day. In actuality, a calendar day is 1/2 night, day, 1/2 night since the calendar day officially begins exactly in the middle of the night, splitting it into 2 days. In essense, when talking about the day there is no problem but refering to one specifing night or the night on a specific calendar day is ambiguous since each calendar day contains 2 nights, and each continuous period of night (a night) is part of 2 different calendar days.

Now I will propose the system that I endorse. What seems logical to me is to have the calendar days break on a day/night boundry. The hebrew calendar does this although it breaks the calendar day on the evening (day to night) border because the bible says that "there was evening and there was morning" listing them in that order. I believe that the other transition is really the optimal place for the break as it is more in line with the human schedule. There are a couple problems with this. First of all, sunrise isn't at the same time every day. Since the length of daylight changes through the year, and the lengthening and shortening happens semetrically about noon, that moves sunrise.

This explains why noon is the basis for the counting of the hours. Since midday is invarient year round (each ones is always 24 hours apart regardless of time of year), it explains why when sundials used to be the main method of keeping time, noon was definded to be a zero. Because, like I said, we consider morning to come before night, this also explains why midnight (the anti-noon) was picked to be the beginning of the day between it and noon as the two logical choices (makes some sense to start the day at hour zero). As an aside, as a programmer, I find it ridulous that we insist on this bizarre counting system that is a combination of zero-based and one-based. If you think about, the system is actually zero-based (why 12am is the first hour, not 1am), except that we call zero, 12. My argument for zero isn't based on that I'm a programmer. When you number things in a list 1 is an acceptable starting place. My argument is mathmatical. On a number line, the origin has to be zero. The hours arent lables, time is a scale. No time since the beginning of the day is 0 hours, just as no distance from a point is 0 meters. Because time is a real (versus integer) quantity, it wouldn't make sense for the first 1/2 hour to be 1:30. If you can't stomache the number zero, I undertand why 12 is the next most logical choice, since it at least follows logically from the previous hour being 11. However, the concession between the two systems to call the hour being 11pm and 1am, 12am and to have 12 techically be before 1 as opposed to after 11, seems more confusing than just forcing people to deal with the number zero.

Back on topic, while I understand why the hours were labelled the way they are, with zero being midday and midnight, despite zero being a logical choice for the beginning of the cycle of the day (as it is the beginning of the cycle of numbers that are the hours), it is not an exceptable time of day to call the beginning of the day. There are 2 solutions to this. One is to call some other time the beginning of the day (despite it not being the beginning of the numbers, which isn't any more abitrary than using 12 to represent zero) or shift the labeling of the hours so that zero can be the beginning of the day at a reasonable time, now that we are not dependant on sundials. Even if we were, we're only talking about an offset, as if you were to abitrarily label the one measurable centerpoint on a sundial as 8 (for example) instead of zero, so that if you count back zero would be where you want the day to start. Although I think the best solution would be if zero were called zero and not 12 and it were set at the time I want the day to start, at this point, our associate between clock time and time of day is strong enough that I don't think it is worth changing the way the hours are labelled and mapped to time of day (even though I think it is bad), but instead to just redefine the start of the day, techincally, to something that will make more sense to people, intuitively. I think this change is possible. The use of timezones and standardized time in the first place is a historically recent event to begin with (less than 200 years), and the amount of times that daylight savings time has been changed in the past 50 years even shows that changing of time isn't a completely unreasonable thing to attempt.

So now the question remains, what time should be the beginning of the day. It can't be defined as sunrise because that changes. In fact, even if it were at the same time every day that scheme might not work because of one argument for midnight. In arguing this I've heard the reasonable claim that if you set the beginning of the day (let's say 6am, as a rough average of sunrise) that it will cause trouble for people who wake up earlier than this. The claim is that while midnight confuses some people in the evening, it is more important that it is the same day when all people (on a fairly normal schedule) wake up, and the trouble in the evening is late enough that even those that haven't gone to bed are likely just relaxing and not doing anything so important that it matters. This brings up a new definition of when the day should start. This indicates that like the international dateline was placed in the middle of an ocean far from where people live so that you wouldn't casually walk across a day difference line, the transition from today to tomorrow should be placed in the middle of when you are asleep so that you won't temporally cross that day difference line. While I argued that our human day cycle is based on daylight and so should time be, it is likely that I feel that way because the intuition is for the time to be based on my human cycle. While basing time on the thing my cycle is based on would be a close estimate, it seems the best thing to do would be to based it on the human cycle itself.

Unfortunately, the human waking cycle is even more variable than the sunrise cycle. Being someone who has pulled his share of all-nighters, I have thought a lot about when does night become morning. When I stay up all night, usually working on a project due the next day or that I just got so involved in I couldn't stop, there is a point where I stop thinking about it being night where I have nothing to do but work and when I start thinking about what I have to get ready to do this approaching day, much as might when I first wake up if I had slept. For me this usually tends to be when the sun starts invading my cave of a room and makes me realize that day is close. We've already rejected sunrise, so that doesn't work, but I think the idea that it is when it starts to "feel" like morning is approaching the right answer, but in the wrong frame of reference. The other answer I had is 4am. I used to think of 4am as the transition point because I had decided it was the latest I could go to bed and still get up at a reasonable time and I basically had the principle that 4am was the point of no return; if I was up past 4, I was better off staying awake. When I got a little older and stopped having morning classes, I found that sleep was important and that even if I didn't get home until 8am but had nothing important to do that day that I should go to sleep. Still I felt that there was something pivotal about 4am.

My new reasoning is that if you want most people's waking hours to fall within only 1 day, you want to bisect the average night. Midnight bisects the physical night, but we need a start of day that bisects the average human night (sleeping period). Since people have all sorts of schedules, it would be impossible to pick a time that works for everyone. However, I think the solution to picking the time that works the best is to pick the one that would effect the fewest people. The result of this should cause an equal number of people for whom the day change would occur in their morning as in their evening (as opposed to midnight which almost exclusively occurs in peoples evening) but the number would be minimal, which midnight certainly is not. The optimal time would be the time which crosses the point where now more people are waking up than are going to bed. Put another way, of all the people awake at that hour (that it would therefore affect) an equal number should be still awake from the previous day as the number of people already awake for the next day. Any earlier and the number of people still awake should be greater; any later and the number of people already awake would be greater. This time would be the one that an equal number of people consider night as consider morning. I believe that 4am is this time. You hear about people who frequently stay up until 3-4am, but staying up until 5am or later is usually something that few people do and then only occasionally. You also hear about people who have to wake up very early to go to work, usually 4-5am being considered early, whereas references to having to wake up before 4am are generally either for special (often travel related) circumstances or for people working strange jobs with odd shifts. Although tests could be done, I intuitively feel that 4am is the optimal time for a day to begin.