Re: [office-formula] Date counting conventions in OOXML

["David A. Wheeler" <dwheeler@dwheeler.com>]

> Does anyone have a good idea what the statement, "If the date is 28 or 29
February, it is adjusted to 30 February" means?  Once you break outside
of the Gregorian calendar and introduce fictitious dates like 30
February, you really need to specify their semantics.  For example, is 30
February an alias for 1 March or 2 March?  Or is it an new intercalary
date interposed between 29 February and 1 March?


The best information source I can point to is the ISDA (International Swaps and Derivatives Association, Inc.) document "EMU and Market Conventions: Recent Developments".
http://www.isda.org/c_and_a/pdf/Euro-swap-memo-ISDA-1998.pdf

I just found that Wikipedia has a wealth of very-interesting references, some of which I hadn't seen:
http://en.wikipedia.org/wiki/Day_count_convention

Of course, we have the continuing problem of not being certain WHICH convention maps to WHICH basis number.  THAT is the real problem; there are many different conventions documented in various places, but it's hard to determine which (if any) are REALLY meant by the various BASIS numbers.  Excel has even changed its calculation results over time, and it's not documented anywhere, so it's very hard to figure out the "intended specification".

--- David A. Wheeler