[Bug 2571] R-090: Questions about the lexical and canonical rep'ns of dateTime

http://www.w3.org/Bugs/Public/show_bug.cgi?id=2571





------- Additional Comments From sandygao@ca.ibm.com  2005-12-09 04:46 -------
A conformance note is only available for duration. We need to considering doing 
the following:
1. Apply a similar note to other types.
2. For duration, provide conformance criteria for other fields than year and 
second. The lex value "P99999999999999999999999999M" is currently allowed, but 
I don't think we want to require mininum conforming processors to support it 
when they are allowed not to support "P10000Y".

For #1, the solution is abviously to apply the note we have for duration to 
dateTime, time, date, gYear, and gYearMonth.

For #2, there are 2 ways to solve it: either we specify limits for each field, 
or we specify limits on the equivalent (month,second) values. That is,

Option A: All minimum conforming processor must support duration values 
(y,m,d,h,mm,ss):
1) y <= 9999
2) m <= 11
3) d <= 30
4) h < 24
5) mm < 60
6) ss < 60
7) ss with at most 3 fractional digits

Option B: All minimum conforming processor must support duration values 
(y,m,d,h,mm,ss):
a. abs(y*12+m) < 120000
b. abs(d*24*3600+h*3600+mm*60+ss) < 31*24*3600
c. ss with at most 3 fractional digits

Received on Friday, 9 December 2005 04:47:09 UTC