| Subject: | Documentation for BC dates |
As I was working on my Perl 6 module Date::Calendar::Hebrew, I have
written some Common Lisp code and some Perl 5 code to convert Hebrew
dates to Gregorian and compare four Perl 5 modules with Reingold's and
Dershowitz's calendrica-3.0.cl in Common Lisp.
The two source files and the result file are annexed to this
RT ticket.
As you can see in the result file, there are differences between
calendrica-3.0.cl and Date::Converter. In my opinion, it is *unfair*
to call these differences "errors". They all involve BC dates. I think
that Date::Converter uses the historian way of dealing with BC dates
while calendrica-3.0.cl uses the mathematician way. On the other hand,
there is a *remote* possibility that Date::Converter is buggy with BC
dates.
Here is a more detailed explanation, ready for cut-and-paste
=head2 Dates before the Epoch
The following explanations apply to the Gregorian calendar, but the
general idea applies to all calendars included in C<Date::Converter>.
There are two ways of dealing with dates before a calendar's epoch,
e.g. 1st January 1 AD for the Gregorian calendar: the mathematicians'
way and the historians' way.
The mathematicians consider that the numbering of the years follows
the natural flow of the integer numbers. Before year 1 you have year 0
and before that, year -1, year -2 and so on. The eve of 1st January 1
AD is 31st December 0. 366 days before 1st January 1 AD is 1st January
0 (remember the leap year rule). And 367 days before 1st January 1 AD
is 31st December -1.
On the other hand, historians do not use a zero year. With their
numbering scheme, the year before 1 AD is 1 BC.
C<Date::Converter> uses the historian numbering. The only caveat is
that "1 BC" is expressed as a numeric value C<-1>, "2 BC" as C<-2> and
so on.
end of copy-paste
I have not checked how Date::Converter deals with the negative years
in other calendars. If they behave differently from the Gregorian
calendar, you will have to make some alterations to my explanation. I
have supposed that the module implements all carendars in similar
ways.
You may also add this other paragraph
=head3 Proleptic Calendar
Some people may argue that the discussion above is pointless because
there are no Gregorian dates before 15th October 1582. The Gregorian
calendar used in this module is the I<proleptic> Gregorian calendar,
that is, a calendar extending the Gregorian calendar into the past
until the BC years, while keeping the Gregorian leap year rule and the
fact that the beginning of the year is the 1st of January, not the
25th March or Easter or some other date.
end of copy-paste
And if my supposition about using the historians' numbering scheme is
wrong, if the processing of negative years is actually a bug, can you
fix the arithmetic in your module please?
And while it is not the place in a bug tracker, I want to add:
-- Many thanks to you for writing "Perl Deep Dive", which helped me
a lot to write Perl 6 modules and Perl 6 code (similar thanks to
Laurent and brian).
-- Many thanks for writing Date::Converter, which was the basis for
the algorithmic part of Date::Calendar::Hebrew.
-- Many thanks for organizing several YAPC::EU conferences, plus
a few days ago PerlCon in Riga.
| Subject: | comp-p5 |
Message body not shown because it is not plain text.
| Subject: | comp-p5-result |
Message body not shown because it is not plain text.
| Subject: | cvt-he-gr.cl |
Message body not shown because it is not plain text.