On Sat, 21 Jun 2003, Joshua Hoblitt wrote: > > It is not clear to me what a value like 2003-06 would mean. Is it "a > > timepoint somewhere in June 2003" (an imprecise datetime), or is it "all > > timepoints in June 2003" (a datetime span)? > > This is why I feel the need for time only objects or at least time only > math objects. Forcing a date to be attached is like adding zero's past > your known precision.
Here's the thing. Yes, the object would contain more precision than the original data, _but_ presumably if you are only exchanging "year and month" data, then you will only look at the year and month of the returned object. It's really hard for me to think of a case where you would not know the expected precision in advance. -dave /*======================= House Absolute Consulting www.houseabsolute.com =======================*/