|
Edward Rice sent this comment about year 2000 apropos an article
I had written
(www.leuf.org/articles/19980206.htm)
Hi -- this is on your web-site and is in error:
> The well-known rule that every 4th year is a leap year,
is modified by a less well-known rule that even centuries are not -- thus
the year 1900 had only 28 days in February. (Any number of calendar calculation
routines still get this wrong!). However, every so often there is an extra
correction of cumulative error to be made, so it was recently determined
and decided that, contrary to rule and expectation, the year 2000 would forego
the expected 29 days in February.
In fact, the "rule" (normal case) is 28.
The first exception is multiples of four, which are 29.
The exception to the exception is multiples of 100, which are 28.
And the /standard/ exception to the exception to the exception is that
multiples of 400 are 29 again.
What I'm trying to say is, the "special" case you cite for the year
2000 is in fact not special at all -- it wasn't decided recently. It's part
of the standard date computations, and will occur again in the year 2400,
and again in 2800.
Regards
Thank you for the feedback. I stand corrected -- had forgotten about
this piece written a year ago, when things were much more confused and local
news was expounding, apropos the formal decision, both versions: leapyear
and not.
The problem with date and time, i.e. calendar standards, is that
most people assume that the rule dictates reality. In fact, the rules must
now and then be tweaked to correspond to measured reality. Our current
leap-year-with-exceptions rule is holding up fairly well, but it is *correct*
that there was in fact a *formal decision* taken relatively recently about
the leap year status of year 2000. I admit that my formulation was perhaps
less clear about that the decision was confirmative, but the 4-century rule
is virtually unknown even among programmers who code date
algorithms.
Whether the years 2400 and 2800 "will" in fact be leap years or
not is an unknown. The current rules predict it, but any number of things
can happen in 4-8 centuries, both physical and with regards to calendar
standards.
The bottom line point made is still valid however, that algorithms
which dictate correct date and time are a risk, since
predictive is not the same as
determinative. Last year's "alignment" of summer time in
the EU is a case in point, since patches had to be distributed to fix the
incorrect automatic adjustments in Windows.
The article referred to has been modified accordingly, with edit markup.
We know the truth of a predictive standard only when the events
become history. Until then the rules can change, and sometimes even the reality.
Tom Syroid comments on
today's story about driving error...
Yes, my friend, you are indeed lucky Mr. Murphy let you off the hook
on this one. Looks like you got caught in HICS -- and lucky for all involved
you realized the error of your ways before any serious do-do starting falling
from the skies.
Your fundamental error, of course, was in ASSUMING you knew the reason
for the stopped car. And we both know that assuming almost always makes an
ASS-out-of-U-and-ME. Simply put, assumption is very ungood in today's complex
world of driving. There is no room for it, cuz there is no room for
error.
In our world of fast cars, decent roads, daydreaming drivers, and
expensive tin, errors in driving judgment are at best expensive and at worst
life threatening.
And just in case you think all this sounds a little condescending
on my part, let me share with you why I knew exactly what was coming in your
story: Because I've BTDT myself. On more than one occasion, I've made the
same assumption you did this morning and lucky for me I didn't hurt anyone
either. Providing, of course, you don't consider the amount of adrenaline
I instantaneously injected into my body harmful in any way.
My experiences over the years has taught me that driving in today's
day and age requires much the same mentality and focus a Special Forces Commando
must bring to his job: There is a time for fast and a time for slow -- know
when to employ each; there is no such thing as too much training or knowledge;
assumption is the mother of all F-ups -- don't do it; don't let yourself
tunnel-vision -- you'll miss the other 95% of what you need to be paying
attention to; and always, always, always ensure you are thinking at least
6 steps ahead of the enemy -- which in this case is everyone and everything
sharing your space (and don't ASSUME that implies just vehicles)
"Special Forces Commando" -- yes, I can relate to that.
Ok, today's mission impossible, should you choose to accept it, is to drive
crosstown in 8 AM traffic.
More about the leapyear thing... Edward Rice added the historic
detail of the origin of the 400 year exception:
The decision was made in 1582 (Pope Gregory XIII), when the Julian
calendar was shifted over to the Gregorian. The calendar shift at that time,
to take into account detail beyond the quarter-day-per-year, handled the
century and 400-multiple years, and will work fine for about 4,000
years.
I was not disputing the origins of the exception rule. What I clearly was
unclear about was that the "recent formal decision" was a confirming one,
not an originating one. Oh well, part of my reply went...
Sorry, but the point I was trying to make is that leapyears as applied
is an arbitrary rule. There is nothing inevitable about a given year being
a leapyear. Rules do change, arbitrarily, given major social upheavals --
history is full of them, and ours is not the only contemporary calendar,
nor the only algorithm for keeping calendar and astronomical timekeeping
in sync.
Bob Thompson made a similar comment:
Well, I suppose it depends on what you mean by "relatively recently."
In fact, it was 1582 when Pope Gregory developed the method we currently
use, including the rule that years that end in two zeros must be evenly divisible
by 400 to be considered leap years.
And I'm not sure what you mean by your point about predictive versus
determinative. The method we use to label the passage of time is purely
conventional, with no basis in physical reality. The time that the earth
takes to rotate once on its axis is certainly known within very small error,
as is the time required for the earth to orbit the sun. Obviously, neither
of those values is "fixed" in any cosmic sense, because each is gradually
creeping away from its current value. In an eon or two, both will be noticeably
different than they are now.
But the method we now use to reconcile the fact that the earth's rotational
and orbital periods do not share a usable common divisor is fixed by
international agreement. So, it is both predictive and determinative in the
sense that anyone since 1582 who cared to check would have known that the
year 2000 was a leap year. It is certainly possible at some point that the
rules will change, either within the existing context or to make the
intercalation even more accurate. For example, everyone could agree that
the years 2001, 2401, 2801, etc. would be leap years rather than 2000, 2400,
2800, etc. But that wouldn't change the intercalation method. Equally, we
could take the existing small error into account (by continuing to use the
existing method, we'll be off by one day by the year 4915) by changing the
rule to consider millennia only if they are evenly divisible by 4000. And
so on...
Arbitrary, yes, but I'd not say no basis in physical reality. We
have however regulated and abstracted it to a precision that outstrips the
original yardsticks. The point is that many people mistake the model for
the reality it models.
The point is also that human history is full of arbitrary re-adjustments
of calendars, restarting year cycles and whatnot. It's possible that in 400-800
years the dominant calendar may look quite different, with other cyclic
adjustments. I.e. the current algorithms are predicitive only within their
context; they alone cannot determine what happens in the future. Rules can
change, yes, both to improve accuracy, and to worsen it, depending on the
regime.
Edward Rice closes with...
Agreed. Russian history is full of calendrical changes, and the French
have done more than their share of messing around with it, also.
It's bad enough that the Antarctic ice cap is apparently melting. I
can only worry about one major disaster at a time!
And Bob Thompson again:
Okay. Sounds reasonable. Regarding arbitrary changes to things with
a basis in reality, the Indiana state legislature actually passed a law that
fixed the value of pi at 3.0.
|
Mail
your comments to:
bo@leuf.com
