Worst problems are those that remain unknown, awaiting on darkness to attack in proper date, and I know *one* of such problems awaiting for unaware people. It is a Y2K cousin problem, also related to year 2000, and probably will attack many Corporations and Government agencies on February 29, 2000.
It is known from astronomy that Earth turns around Sun at every 365.2422454... days. This 0.2422... fractional part is the reason for leap-years (those years having an extra day on February). If you multiply this odd number x 4 will notice that every 4 years Earth will go 1460.9689.. days, or ALMOST 1461 days. Well, if astronomers just maintain one year with 365 days for rounding purposes, every 4 years will obtain 1460 days, and Earth will be 1 days ahead than foreseen. So, where is the extra day?. To solve this problem, Jules Cesar, roman emperor, established some 2045 years ago the "bissextile year" (what a clever people where those romans!), or "add from nothing" an extra day, 29, to February. This way, at every 4 years Earth would be 1461 days ahead, and because this, it was named "Julian Calendar".
BUT 1 is NOT equal to 0.9689... and by year 1582 this tiny difference had accumulated to some 12 days more than Earth really performed (1 - 0.96... = 0.031.. , multiply this x 1600 and divide by 4 = 12.40... days)
This shifting was enough for people begin to notice seasons was becoming on wrong places, and that year (1582), Pope Gregory XIII established the "Gregorian calendar", cutting off 10 days (yes, after 11:59:59 PM October 4 1582 coming 00:00:00 AM October 15 1582 - what a killing year!).
That's the reason you won't find docs dated for example "October 9, 1582" on Vatican documents. Yes, no newspapers, HTM pages or account payable checks had those kind of dates, and I suspect also credit-cards having debts on those dates were cancelled, as well as trips and travels via Caravelles, Horse cars, subway and air planes. Some countries whith Protestant religion (England...), held the change until 1752, when cutting off 11 days on September 1752 (on those countries life became frozen from Set 3 to Set 13 1752)
But Pope Gregory performed another correction also, and now entering our killer problem. To avoid future crazy shifts, he established that every 100 years Feb. 29 would be BANNED, that's to say, there's NOT February 29 1900 nor February 29 2100, even those being 4-divisible years (please check your "cal" unix and MS-DOS commands, UNIX is ok, some MS-DOS are WRONG!), And still more, every 400 years, Feb. 29 INDEED would exist, even these being 100-divisible years.
If you compute all those operations, will notice what Gregory XIII was trying is to adjust in some way the 0.2422... fractional part, because 1/4 - 1/100 + 1/400 = 0.2425 (and there's more adjusts still to come I can't exactly remember now, but every 3 or 4 thousand years Feb. 29 also will be banned, please remember me to post again this message on 5000 A.C).
You noticed I said "some MS-DOS" cal command are wrong (yes, some of them INDEED show Feb. 29 1900) ; why? , Because programmers usually uses a math approximation formula to calculate "days of week", or to solve problems like "what day is Jul. 31 1999?, Answer: Saturday" into their programs and systems. Unix programmers used a *good* enough formula, some MS-DOS programmers used a *bad* enough formula to "cal" (calendar) command, because some programmers were tricked by the idea that " if 100 years are a LOT of years ahead, then 400 years are still more a LOT of years ahead, nothing to do with *my* system". (Please note MS-DOS not having "cal" command developed by Microsoft, but rather some independent programmers developed shareware programs to emulate the Unix "cal" command).
Those people, at the moment of writing code, just turned heads to partners screaming "hey people, does anybody know what's the exact formula for fitting days into year?", and since Encarta and other electronic encyclopedias only came alive from 1990 on now, you may imagine that info was not so easy to get. Then, a lot of programmers and system analysts just heard from partners "...hey John, I think a good approximation is 365.24, use this formula instead, it works flawless into my accounting system..", and that was true, those formulas had performed very well on computer systems for last 300 years. This way, many account payable, saving accounts, stock positions, travelling agencies, CIA agencies, White House agendas, have probably into some thousand million codes, hidden the terrible 365.24 "good approximation" number.
Well, fact is problem is turning the corner, because 2000 *IS* a 400-divisible year, and for all those thousand wrong programs, February 29 2000 WILL NOT exist after all because 365.24 approximation is just enough to cut the February 29 day from all 100-divisible years, and 2000 is AGAIN a 100-divisible year, so it will, that's to say, for all those Bank Corporations, Government agencies and Travelling agencies, after MONDAY 11:59:59 PM February 28 2000, surely will come TUESDAY 00:00:00 AM March 01 2000.
If cutting some 10 days by Gregory XIII in 1582 was a big problem, you might imagine what would be cutting just ONE day on 2000, where most people around the world are regulated by computers. Worst part is that all those Corporations and Companies couldn't continue with their operations on following days UNTIL correcting totally the flaw, because even managers and directors ordering to set back the clock of computers next morning, even so computers will persist informing that March 1st is TUESDAY instead correct day: WEDNESDAY. What a nightmare!
Is this really a true menace? If you are in doubt maybe you should check HERE or else here and here and imagine some top range financial institution discovering this on Feb 28 2000.
Into my novel "The Song of the Swan, Part II", I talk a little on some Y2K flaws, and how hackers use this flaw to penetrate computers robbing a secret hidden extraterrestrial message info recorded by government agencies.
My suggest for all corporations and government agencies is they should perform systematical checks also against Feb. 29 2000, and not only against Jan 01 2000, because this flaw could be a lot more dangerous than the "Y2K" well known problem. See how Argonne National Laboratory performed an Operating Systems checking for Feb. 29 2000 (this check just assure OS is properly working, nothing it says about customer applications and programs).
To know a little more about Y2K problem itself click here
To know a about Y2K survival kit click here
To go home click HERE
LINKS (note I didn't check accuracy for links info, but just references)
All you want to know about calendars from Larry Freeman pages.
Document about "Y2K compliance" PA96014 from GTE Technology.
Here the CISCO Year 2000 Compliance Testing
How to compute Day of Week using computer? (try this one also), or even just using your mind?
How does it look source code for "Calendar" command?
Java app. to calculate date conversion (Greg. x Julian, Jewish and French) Here
Unix Shell scripts for date conversion and difference in dates Here
