On Tue, 2006-10-10 at 20:39 +0100, Benoit Xhenseval wrote:
> Dear Joda People, Paul and Julen,
> 
> Thanks for your comments, it is appreciated.
> 
> Steve, thanks for posting announcement to the list.
> 
> Paul, I agree with you, this would be a great addition to the Kit.  It is
> not explicit on the interface when the last 'holiday' would fall and whether
> this may impact a calculation or not.  We have several options to consider,
> we can elaborate here a bit if you wish:
> 
> Let's think about requirements:
> - should the calculation fail and an exception be thrown if the date
> calculated is beyond the last holiday?

I don't think that last holiday (or first holiday) is necessarily the
right idea. For example a fiscal year calendar which covered the period
2006-09-01 to 2007-08-30 in the USA might be valid for the entire period
even though first holiday is the 3rd of September and last is the 4th of
July (you get nearly 3 months for free without any explicit values in
the table).  Same for a calendar Year, just because the last holiday
listed is Boxing Day (to use a UK example), doesn't mean the list of
holidays is not good for the entire year 1/1 through 12/31.

> - May be this could be a setting on the calculator (like
> setEnforceHolidayBoundaries(true) or setStrict(true)?

So what use case are you thinking of which wouldn't want to let the 
developer know that applications are asking outside of the valid range
for the list of holidays?

> - when it comes to combining n Calculators (you are not limited to 1), we
> took the easy approach (sorry the XP approach) 

Don't wory, I'm all for an XP approach, but I think that valid range of
a DateCalulator is the next idea to incorporate.

> and created a new set of
> holidays based on the combined calculators'  This works nicely but would not
> allow the detection of an issue at date calculation-time (it does not
> remember that calculator1 had only 2006 and calculator2 up to 2007). But
> that could be added as an extra internal attribute during the combination
> (we would keep the lowest denominator but only if a calculator had 'some'
> holidays, not if it is empty).

Sure an extra internal which provides the complete _narrowest_ range of
all combined DateCalculators (that is one calculator might have all
Christmases for a very long time, but another might only list the 
bank holidays for next year, so the combination of the two is
just good for next year.

Just my $0.02 worth.


> 
> What do you think?

HTH,

-Paul



-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Joda-interest mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/joda-interest

Reply via email to