Mark Reed wrote:
>On Sun, May 18, 2008 at 1:13 PM, ROGER MILLS <[log in to unmask]> wrote:
> > Their leap days fall nearest whichever equinox/solstice is most out of
> > whack.
>
>That's an odd combination, then.

True, it's arbitrary.

You have a purely arithmetic rule
>about *when* a leap year occurs (years 4, 10, and 17 of a 19-year
>cycle),

I don't think so: according to my calc. 700 - 706 - 713 - 719 - 725 - 732 - 
738 etc. were leap years, so it goes 6-7-6

but then you place the leap day within that year according to
>astronomical observation.  Which makes developing a Cindu converter in
>software very hard, by the way!  But the point is it would be more
>consistent, at least to human logic, if you used either astronomy or
>arithmetic for both, i.e. "a leap year occurs whenever any of the
>equinoxes or solstices is off by more than a day" .

That would be every 6.333 years.
>
>On Earth, that rule would be problematic because the equinoxes and
>solstices move relative to each other in ways that make it impossible
>to correct for all of them; you have to pick one to go by.
>
>Alternatively, and this is what I would recommend as a programmer
>implementing the conversion, you could fix the leap day according to a
>simple rule.  e.g. "the leap day is always the day before the spring
>equinox" or

"__each leap day occurs a quarter later than the last leap
>day" (the specific date chosen according to your waiting/dancing rule,
>of course).__

I like that better than having it be a fixed day.  (NB I am not skilled at 
this sort of thing!!! I'm still digesting your long previous post..........)
>