Re: [xml-dev] Incompleteness of Duration

[Michael Kay <mike@saxonica.com>]

On 25/04/2012 16:40, John Cowan wrote:
> Greg Hunt scripsit:
>
>> To the extent that ISO is a culture all to itself (the ISO definition of a
>> week does not align with traditional Christian, Islamic or Jewish calendars
>> but does seem to be well suited for the Western working week), Mike's point
>> about the start, definition and numbering of weeks seems to be reinforced
>> by your comment.
> Well, no doubt.  But the fact that YYYY-MM-DD is a relatively uncommon
> way of writing dates (mostly found in East Asia) didn't stop ISO from
> standardizing it or XML Schema datatypes from adopting it.  So "Ways of
> conceptualizing the week are diverse" is not an argument against accepting
> the ISO way, any more than "Ways of writing the date are diverse" is.
The use of a numeric day, month, and year to represent a day in the 
calendar is very common around the world, and all that ISO did was to 
define a machine-readable notation for expressing the day, month, and 
year for use in software-to-software communication. Wider adoption of 
that format for human consumption came later, and was not part of the 
original intent of the ISO standard.

As for week numbers, as I said before, it's all a question of how you 
decide where to draw the line. As with proposals to add things to the 
XPath function library it's very hard to find a convincing objective 
test for what should go in and what shouldn't. If you shout loud enough, 
and give the impression that you're not the only person shouting, then 
people will eventually put it in to keep you quiet - lots of 
inappropriate things are in the XSD specification for that reason.

Michael Kay
Saxonica