[wms-dev] Changes in ISO 8601:2000 revision

[Jeff de La Beaujardiere]

Allan Doyle writes:
>   http://www.merlyn.demon.co.uk/datefmts.htm#8601
> points to some links about the ISO spec, including a 2000-12-15 draft in
> PDF.

Thank you!

There are some changes of note in the 2000 version:

* The year 0000 is defined to be the year before 1 AD
  (equivalent to 1 BC in Dionysius Exiguus's unfortunate
  numbering scheme).  A public redefining of 1BC as 0000,
  and subtracting 1 from the absolute value of all other "BC" years,
  would allow us to avoid silly arguments at the start of the next
  century.

* A leading '-' sign can now be used to denote years before 0000.
  However, leading '-' signs can still also denote omitted portions of a
  truncated representation as used in GeoParsing services (e.g., the
  16th of the month at 10 am is '---16T10'), so each "information
  community" must determine exactly what formatting options it will use.
  WMS 1.1.0 currently handles negative years by using the prefix 'B':
  27BC = B0027.  This would become -0026 in the revised ISO standard.

* Years larger than 4 digits are allowed, both positive and
  negative.  WMS 1.1.0 currently handles geologic times
  using the prefixes K, M and G for 10**3, 10**6 and 10**9
  years ago (following geologists' convention ka, Ma, Ga).
  Large negative years could theoretically be adopted instead:
  Jurassic period (150 Ma) = -150000000 instead of M150.
  This would add bulk, but avoid using non-ISO prefixes.

Thankfully, because WMS 1.1.0 adopted the punctuated form
yyyy-mm-ddThh:mm:ss.sss, we could safely add a '-' and/or extra digits
to the year without ambiguity: -1121 = 1,121 years before 0000, but
--11-21 = Nov. 21 of unspecified century and year; 20010716 = year
20,010,716 "AD", 2001-07-16 = July 16 of 2001.

-Jeff DLB