[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Scheme-reports] Date/time package

John J Foerch scripsit:

> Okay, let's suppose an application working with historical astronomical
> data.  The tradition in that field, since Kepler I understand, is to use
> a calendar which is Julian up to 1582 has the Gregorian reform in 1582,
> and *has* a year zero.  How would this proposal accomodate that
> application?

The best I can think of is to make a compound calendar and then hack your
own year procedure which adds 1 to all negative years.  Alternatively, you
could use implementation-specific methods to write your own base calendar.

Similarly, the system can't handle the unholy mess that Sweden made of
its calendar.  It was Julian until 1699, but then 1700 was not a leap
year, leaving Sweden with a unique calendar.  To return to the Julian
calendar, 1712 was a double leap year (it had February 30).  Finally the
Gregorian transition was made in 1753.

Nothing's perfect, and the more knobs we add for special cases, the harder
the system becomes to use.  I think we are just about at the limit now.

If you understand,                      John Cowan
   things are just as they are;         http://www.ccil.org/~cowan
if you do not understand,               cowan@x
   things are just as they are.

Scheme-reports mailing list