Patrik:
Good reference. This is impressive work, even if it is in support of that evil data schema work ;-)
A quick review seemed to show that the work did not cover the WEEK component used in ISO8601. It only covered YEAR, MONTH, DAY and then the time components.
In addition, the work too the narrow view on timezones that it was just a UTC offset. We know that timezones involve more than just a scalar offset. There is the issue of DST behavior too.
But still a good reference. I certainly bookmarked it. Thanks!
-- Frank
People should read the following document before comparing days and hours (or rather, before starting comparing two intervals):
http://www.macchiato.com/unicode/timeIntervals.htm