On Thu, 2003-01-23 at 16:35, Dave Rolsky wrote: > On Wed, 22 Jan 2003, David H. Lynch Jr. wrote:
> You're confusing phsyical with logical. A given DBMS may _store_ dates as > julian days (or MJD, or whatever) and you'd never know, since you only > deal with the representation of the date (as a Gregorian date). > > If you're suggesting that JD would be a good representation, I'd disagree. > Actually I beleive the SQL standard screwed up, and basically confused data with presentation. The display format of the data, is purely a presentation issue. SQL should have specified that Dates were stored transfered etc. as julian Day numbers. It then could have specified that the implimentor could translate whatever the user might have typed into julian day numbers based on the appropriate institutional, governmental, or local rules. Further one of the things I never fully comprehended regarding the Y2K problem was that a 3 byte julian day value can handle all of recorded history (46K years) - it can handle most of it even encoded as BCD (2.7K years). Even a 2 byte Jdate with 1/1/1900 as 0 can cover almost 200 years before overflowing. This is much the same as the difference between numbers and currency, although more pronounced. The fixed point value 10000.00 is the same regardless of what part of the planet you are on. However the presentation of that value may be unique to one part of the planet or another, and may vary even more widely if that value represents currency. > > -dave > > /*======================= > House Absolute Consulting > www.houseabsolute.com > =======================*/ -- David H. Lynch Jr. <[EMAIL PROTECTED]> David Lynch & Associates
