ISO 8601 v2

[Markus G. Kuhn]

I have just received a copy of the document ISO/TC 154/WG 8601/101
"Proposal for ISO 8601 - version 2", 1996-05-02, from the Nederlands
Normalisatie-instituut (NNI).

Here is a very brief summary of the suggested changes:

- a new section explaining the difference between "time-point" and
  "period (time-interval)", as well as "duration". "This standard
  does not specify what entities should be considered time-points
  and what entities may be periods or time-intervals. The decision,
  whether for instance 'April 1985' specifies a time point or a
  time-interval is left to the users of the information (because this
  decision depends upon its use.)".

- The units of time used and the zero-points of counting schemes are
  now explicitely defined:

  - second: as in ISO 31-1
  - minute: time intervall of 60 seconds
  - hour: 60 minutes
  - day: 24 hours
  - week: 7 days
  - month: approx. 30 days
  - year: 12 consecutive months, considerd to approximate
    the time-interval required for one revolution of the
    earth around the sun

  "The Gregorian calendar uses a starting point which assigns the year
  number 1875 to the year in which the Convention du Metre was signed
  at Paris. All years numbered according to this convention have,
  by definition, one specific (calendar) year number. [...]"

  Ok, so definitely no Julian calendar, but still no word about
  years before 0001.

  The lengths of the 12 months are now listed, together with the day of
  year numers that every months covers in a leap or non-leap year.

- ISO/TC 154 has received a number of suggestions to improve the
  Gregorian calendar in both the lengths of months as well as the
  location and occurance of leap days. "However, ISO/TC 154 does not
  feel qualified to propose a reorganization of the Gregorian
  calendar." (I guess the world has already enough problems with
  the year 2000, and considering the huge investment in almost
  unchangeable computer software, a calendar reforms today seems
  extremely unlikely.)

- the ---D format has been removed.

- the proposal explicitely does not address leap seconds,
  the allowed second range is still 00 to 59 :-(

- the ugly T is still there, although France has critisised this
  separator and the whole idea of considering date and time and time zone
  to be separate fields.

- There is a new annex C "temporal relationships", which defines 13
  terms in order to represent the temporal relationship of an event
  with a time-period. This was copied from a CEN Medical Informatics
  working group proposal. (Make sure you use a monospaced font to
  display the following!)


  Reference time interval:                  ---------------->

  AT                                        ---------------->
  
  BEFORE                       ------->
  AFTER                                                         ------->

  DURING                                         ----->
  INCLUDES                     ---------------------------------------->

  CO-CONTINUES                              --------------------------->
  CO-PRECEDES                  ----------------------------->

  CO-STARTS                                 ---------->
  CO-ENDS                                        ----------->


WARNING: This was an incomplete summary of a 14 page document, which
is only a first draft that is expected to be superceeded by a second
draft in a few weeks.

Markus

-- 
Markus G. Kuhn, Computer Science grad student, Purdue
University, Indiana, USA -- email: kuhn at cs.purdue.edu