RE: [office-formula] Year 1900 bug

["Dennis E. Hamilton" <dennis.hamilton@acm.org>]

  1. The reason it is called a bug is that the dominant implementation
(going back to at least Lotus 1-2-3) allows 1900-02-29 to be an admissable
date.

  2. Another way of looking at it is that there is the presumption that
ordinal days, given an origin date, are in one-to-one correspondence with
calendar dates.  That is, if two dates are adjacent in the calendar
progression of dates, the difference between their corresponding ordinal-day
numbers will be 1.  The bug can be seen as something that violates that
expectation.

As you demonstrated, there are ways for setting the origin date and calendar
mapping so that (1) doesn't happen, at the cost of (2) not holding.  

As Rob points out, we should not be talking about bugs in the OpenFormula
specification.  We do need to allow some precise way to address ordinal-day
calendar correspondences that accomodate the present reality.

-----Original Message-----
From: Andreas J. Guelzow [mailto:aguelzow@pyrshep.ca] 
Sent: Sunday, January 24, 2010 22:34
To: office-formula@lists.oasis-open.org
Subject: RE: [office-formula] Year 1900 bug

On Sun, 2010-01-24 at 15:16 -0800, Dennis E. Hamilton wrote:
> Your proposed wording is certainly an improvement.

I am not sure about that. But perhaps I don't even understand what hte
"Year 1900 bug" is all about, since we do not specify the exact serial
numbers anyways.

[ ... ]