RE: [office] ODF 1.2 draft 7 - table chapter

["Dennis E. Hamilton" <dennis.hamilton@acm.org>]

I am a calendar and date-time algorithm junky, so this exchange provoked a
little research on my part.

I don't have any insight about the non-Gregorian calendars, but it would be
great to provided references to authoritative sources that can be located
and used.

For the Gregorian calendar,

It strikes me that IS 29500 punts on the rules for Gregorian Calendars as
much as ODF does, although ISO 8601 does specify the Gregorian calendar in
an indirect way.  If you know the rule for leap years plus the standard
calendar, all you need to do to make a Gregorian Calendar is know the day of
the week of January 1 for that year.  (It is a Saturday for every Gregorian
year whose number is a multiple of 400.)  There is sufficient information to
deduce reliable rules from that much.  ISO 8601 is more about recording the
dates than computing them. ISO 8601 is rather vague around specification of
durations without known context when there are units larger than days.
There are two old (ISO 30 and ISO 31) specifications that are referenced for
the fundamentals and maybe they help with regard to time sequences.

I have no idea where one might find a better, authoritative modern source on
the Gregorian Calendar.  Someone must have something (although the
experience with time intervals in financial calculations gives me doubt
where I previously had none).  My sparse collection of ISO standards refers
far back to ISO 2014 (writing calendar dates in numeric form), ISO 3307
(representations of time of day) and ISO 4031 (representation of local time
differentials), etc.  None of these have to deal with the fine details of
date-time comparisons, calculations, day-of-week determination, etc., and
8601 supplants them anyhow.  Methods are neither specified nor referenced in
ISO 8601.

Oddly, the treatment in Wikipedia is valuable (and the External links reach
out to a downloadable copy of ISO 8601:2004):
http://en.wikipedia.org/wiki/ISO_8601

The treatment of the Gregorian Calendar is also intriguing:  
http://en.wikipedia.org/wiki/Gregorian_calendar 

Every Gregorian year number divisible by 400 has the same calendar and
January 1 is a Saturday.  Also, there is a 400-year cycle (natch) but there
are actually only 14 different calendars for Gregorian years and you can
pick the right one using anything equivalent to the Doomsday algorithm:
Figure out what the day of the week for January 1 is and also determine
whether February has 28 or 29 days. There are many little arithmetic
shortcuts that can be used. 
  

 - Dennis

PS: What we probably don't care about so much, but should probably be
accounted for in an authoritative source, is the difference between what the
Gregorian date would have been for a relatively-remote past/future day (the
so-called proleptic Gregorian calendar in the case of past dates in places
before the Gregorian calendar was in use) and the date according to the
calendar system that was/will-be actually in use.  (If we are talking about
dates as recorded at the time, we must be careful to avoid confusing them
with what the extrapolated Gregorian Calendar date turns out to be,
however.) There are, of course, non-Gregorian calendars in use to this day,
and converting among them is important.

PPS: If we consider *time*, not just date, there are also those pesky
time-adjustment events that happen from time to time.  I suspect that we
don't care about those, but astrophysics scientist might have some
information on the matter.  Perhaps on the matter of time there is more
interest in the noise around time zones and various local-time adjustments
(e.g., summer time and daylight savings times) in understanding the sequence
and synchronicity of events in time.  Also an interesting challenge, but
maybe not a high priority among those where we may need more precision of
specification.  ISO 8601 indicates how to record everything that comes up in
these contexts, but it doesn't say how to find them.

-----Original Message-----
From: Patrick Durusau [mailto:patrick@durusau.net] 
http://lists.oasis-open.org/archives/office/200810/msg00168.html  
Sent: Thursday, October 30, 2008 10:52
To: office@lists.oasis-open.org
Subject: Re: [office] ODF 1.2 draft 7 - table chapter

Eike,

Just quickly but looking at ST_CalendarType (Calendar Types) in ISO/IEC 
29500 I see:

gregorian (Gregorian) Specifies that the Gregorian calendar, as defined 
in ISO
8601, shall be used. This calendar should be localized
into the appropriate language.

[ ... ]

hebrew (Hebrew) Specifies that the Hebrew lunar calendar, as described
by the Gauss formula for Passover [CITATION] and The
Complete Restatement of Oral Law (Mishneh Torah),
shall be used.

hijri (Hijri) Specifies that the Hijri lunar calendar, as described by
the Kingdom of Saudi Arabia, Ministry of Islamic
Affairs, Endowments, Da‘wah and Guidance, shall be
used.

japan (Japanese Emperor Era) Specifies that the Japanese Emperor Era 
calendar, as
described by Japanese Industrial Standard JIS X 0301,
shall be used.

korea (Korean Tangun Era) Specifies that the Korean Tangun Era calendar, as
described by Korean Law Enactment No. 4, shall be
used.

none (No Calendar Type) Specifies that no calendar should be used.

saka (Saka Era) Specifies that the Saka Era calendar, as described by
the Calendar Reform Committee of India, as part of
the Indian Ephemeris and Nautical Almanac, shall be
used.

taiwan (Taiwan) Specifies that the Taiwanese calendar, as defined by
the Chinese National Standard CNS 7648, shall be
used.

thai (Thai) Specifies that the Thai calendar, as defined by the
Royal Decree of H.M. King Vajiravudh (Rama VI) in
Royal Gazette B. E. 2456 (1913 A.D.) and by the decree
of Prime Minister Phibunsongkhram (1941 A.D.) to
start the year on the Gregorian January 1 and to map
year zero to Gregorian year 543 B.C., shall be used.

[ ... ]

PS: Which raises the interesting issue of what we should do if we find 
one or more of these definitions sufficient? Should we simply cite the 
existing definition? Or for that matter, do we really need to re-define 
Add? Seems like one definition may be enough. Assuming they are 
semantically equivalent. May not be so I am not making a claim that they 
are. But I do think we need to look to see.


Eike Rathke wrote:
http://lists.oasis-open.org/archives/office/200810/msg00166.html
[ ... ]
>
> Btw, regarding definitions of various calendars it seems to be hard to
> find normative references for any of them. You mentioned that ISO has
> a standard about the Gregorian calendar, I found only ISO 8601 that
> seems to specify that it is the calendar as specified by Pope Gregory
> XIII in 1582, and a few dates as reference points such as the metric
> convention being signed on 1875-05-20 and 2000-01-01 being Saturday are
> defined. Is that what you were referring, or is there something else?
> I don't have a copy of the full standard at hand.
>
>   Eike
[ ... ]