Re: XHTML 2.0 <datetime> element proposal

> [Original Message]
> From: Lachlan Hunt <lhunt07@postoffice.csu.edu.au>
>
> Hi,
>   After researching this topic some more, I had a thought, that although 
> the ISO 8601 date format is completely acceptable for most of the needs 
> of writing dates and times in documents, and thus a datetime tag such as:
> <datetime datetime="2003-11-03T00:45:00+10:00">3 November 2003, 
> 00:45</datetime>
> would be fine, however, there are some limitations to this format.

<snip - listing of several non-Gregorian calendars>

>   Most of these calendars mentioned above have differ greatly in the 
> length of one year, and thus the ISO format cannot represent any of 
> these accurately because it is based on the Gregorian calendar only.  
> Not only this, but I do not believe that the ISO calendar allows for 
> both CE and BCE (also known as AD and BC respectively) years.  eg.  How 
> do you write the year 10BCE in ISO format?  please correct me if I'm 
> wrong about this.

The initial version of the standard didn't allow for years outside the range
1AD to 9999AD, however later versions (it was added in either the 1998
or the 2000 revision) do allow for it.  The expanded year format calls for
either a leading + or - followed by four or more digits representing the
year
as agreed upon by the users of this data.  The exact number of added
digits is left to individual applications and according to the standard
should
be chosen to avoid confusion with any of the truncated representations used.
Since XML doesn't use any of the truncated representations (all six of the
formats sanctioned by http://www.w3.org/TR/NOTE-datetime call for using
a complete representation)  simply allowing the year in an XML datetime 
to be represented by either four digits, or either a + or - sign followed
by four
or more digits would allow for such dates to be represented.  (Note:
The year 0 in ISO 8601 corresponds to -1BC, the year -1 to -2 BC, etc.)

>   IMO, creating a system in XML that can represent all these calendars 
> is *out of the scope of XHTML*, and would require the creation of a new 
> XML application, perhaps known as XDate (or XDateTime), which could then 
> be integrated into XHTML, or any other XML document, using an 
> appropriate namespace.
>
>   This would be useful for any computer systems that require to 
> represent dates in other calendars, such as astronomy systems based on 
> 'star Dates'.  It may, of course, be possible to write applications so 
> that they automatically convert ISO 8601 and any other calendar, however 
> this may be inconvenient and take valuable processing time, especially 
> for real time applications that need to work and process data quickly.

I totally disagree here.  First off, which calendars get included?
There are literally hundreds of calendars out there, most of which are
either obsolete or have interest to but a very small community.  Deciding
on where to make the cutoff for a generic standard is likely to prove
contentious, and if we try to include them all, it will be an ever
revising and ever more bloated standard.  Aside from revising
http://www.w3.org/TR/NOTE-datetime to incorporate the expanded
representations of the year that were introduced in revisions to the
ISO standard after that note was written and to perhaps incorporate
a few extra formats using complete or expanded representations,
 I see no need to alter how XML handles time.

Some of these formats such as:
2003-306 (ordinal day)
2003-W44-5 (day of the week and Week of the year)
don't provide any additional specificity, it would be nice and
unambiguous to have.

It would be nice if other time precisions such as:
2003-W44 (week of the year)
2003-11-02T17Z (time precise to an hour)
that don't have much use but would give some added flexibility
were added.

It also would be appreciated if the duration and/or the time
interval vales were specified in NOTE-datetime or a related note.

> [1] http://www.geocities.com/calendopaedia/

Received on Sunday, 2 November 2003 12:24:42 UTC