Re: Compute Date Duration

>Scaliger number 1721424 equates to 0000-12-30 (which is a nonsense date 
>because there was no year 0--hence all the confusion with the morons 
>who think a new century starts when the zeroes tick over). 1721424 
>should be a BC date but CDD converts it to 0000-12-31 when using an SAA 
>Calendar.

Well, yes and no.  Yes, because not having a year zero and counting from 
one reflects history better, because zero was not really understood for 
some time, but no because the history of calendars, especially in the West 
is, as you yourself note, an unholy mess.

Why, given all the oddball fates and (the leap years in the Caesarian 
reign that someone gave is a particularly good example), should we care a 
fig to propagate this particular error of "when the centuries ended," 
entirely a product of ignorance about zero, if we don't want to?

Around the year 2000, when the discussion about this was at its height, 
someone who apparently knows about all these quirks was asked about the 
year zero and when the last century "really" ended.

His response put it all in perspective.  He said something very like:  "We 
can do whatever suits us.  No one actually _living_ in the year zero knew 
it at the time."

And, I would think MI assembler programmers, who daily understand the 
value of counting from zero, would seize the day here and allow, at least 
computationally, a year zero.

And then, there's even this:  Until the Gregorian calendar, the new year 
began on March 25 (the spring solstice in the Julian).  That's why 
"October / November / December" (8th, 9th, 10th month) are named as they 
are.  So, at any rate, our "new year" celebration is off nearly an entire 
season to start with compared to what people did in "1 AD".  Yet, no one 
made a fuss about that!

--

Larry Loen, Senior Programmer, iSeries i5/OS Scaling Analysis
Speaking on my own, of course