Re: DOCBOOK: RFE: Date Format

[kmself at ix dot netcom dot com]

on Sun, Jan 07, 2001 at 11:24:04PM +1030, David Lloyd (lloy0076@rebel.net.au) wrote:

> > So I suggest the valid value for the year be any non-zero interger
> > (negative values to be treated as B.C.E. dates).
> 
> An alternative is that there is a setting:
> 
> <era>
> </era>
> 
> Which could take a default value of AD. This has a big advantage because
> it takes into account:
> 
> * Buddhist's take a different year as "Year 1"
> * Islamic people take a different year as "Year 1"
> 
> And we avoid that cultural war...

Wouldn't that be a role of date?

There are several problems with calendars in general (sorry, prior life
as a data analyst dealing with time and dates).

  - Calendars are an artificial construct.  There are the odd
    inconsistancies which vary by time and place, as well as calendar
    system adopted.  While time and elapsed time are well defined, the
    choice of mileposts and start dates is in fact arbitrary.

  - There's no agreement on what specific date applies in the past --
    Russia's "October Revolution" was due to a late adoption of the
    Gregorian calendar (most of the rest of the world places it in
    November).  In different locales, adoption occured any time between
    1752 and the mid 20th century, with England and the United States
    picking it up in 1792, IIRC.

  - There are other calendars:  the Jewish, Chinese, Mayan, Roman, and
    Sumerian, among others.

  - If you really want to argue idiosyncracies, there are calendars
    which don't follow the Gregorian convention of twelve months,
    or seven-day weeks, let alone start-of-year date (December was the
    tenth month of the Roman calendar, starting in our March), or even,
    in cases, the convention of 365± days.  France, following the
    revolution, experimented with a ten-day week (the "décade"), from
    1793 to 1805.  Soviet Russia's Nepreryvka, was a five-day
    "uninterrupted production week", implemented from 1929 to 1931, then
    a six day week from 1931 to 1940, ending finally in June of 1940
    then the seven-day week was reinstated.

  - DocBook probably doesn't need to be concerned with arithmetic
    accuracy in representing dates.  Managing to localize dates to
    specific calendars is probably an excessive goal.  Most calendars do
    share the common elements of week, month, and year.  However, these
    elements may not follow the same sequence, making numbering schema
    inconsistant.

A general discussion of calendars and their history is found in Eviatar
Zerubavel's _The Seven Day Circle:  the history and meaning of the
week_, University of Chicago Press, (c) 1985, ISBN: 0-226-98165-7.

The suggestion to have day, month, and year elements would require 
date roles of both "era" and "calendar" in order to be fully
generalized.  Defaults could be "CE" and "Gregorian", and stylesheets
could translate formats into appropriate display formats.  

The other straighforward alternative is that dates in DocBook be assumed
to be Gregorian, CE, with other variants simply expressed as literals in
text.

-- 
Karsten M. Self <kmself@ix.netcom.com>    http://kmself.home.netcom.com/
  What part of "Gestalt" don't you understand?      There is no K5 cabal
   http://gestalt-system.sourceforge.net/        http://www.kuro5hin.org

PGP signature