RE: SAG-FO-02 follow-up: Durations

Firstly, I join the regret expressed in
http://lists.w3.org/Archives/Public/public-qt-comments/2004Jan/0408.html.
As I pointed out in
http://lists.w3.org/Archives/Public/public-qt-comments/2003Dec/0283.html,
I favoured a scheme of operations like:

xs:gYearMonth + n
xs:date + n
xs:dateTime + x

to add months, days, or seconds, respectively.

Secondly, now that a decision has been taken by the
WG, only one question from my previous statement I
would like to ask (again): Is it a good idea to

add xs:dateTime + xdt:yearMonthDuration?

The pinning algorithm specified in
http://www.w3.org/TR/xmlschema-2/#adding-durations-to-dateTimes
makes 2003-03-31 + P1M = 2003-04-30. This rule is only
one possible result and it is highly depending on the
use case which result would actually be expected. Some
might, e.g. think of the result 2003-05-01.

In my opinion, a stylesheet programmer who really
faces a use case which necessitates the addition of a
number of months to a date should (have to)
consciously decide on the algorithm to treat days near
the edge of the month. I, too, doubt that we should
encourage coding without thinking. (Compare the
concluding statement in the otherwise unrelated
message
http://lists.w3.org/Archives/Public/public-qt-comments/2004Feb/0102.html).
Thus, I would only foresee an operation to

add xs:gYearMonth + xdt:yearMonthDuration,

which does not prejudge the treatment of days on the
edge of months in potentially unintended ways.

Regards,

Bernhard

--- END ---




	
	
		
___________________________________________________________
BT Yahoo! Broadband - Free modem offer, sign up online today and save £80 http://btyahoo.yahoo.co.uk

Received on Tuesday, 10 February 2004 06:25:14 UTC