Re: Small cleanup in fn-financial.c

["Andreas J. Guelzow" <aguelzow taliesin ca>]

Jody Goldberg wrote:

> > I don't think there's much point trying to fluff the functions up too
> > much; despite their attempts to appear otherwise they are really
> > U.S.-specific, and a proper bond analytics implementation would not
> > have an exposed interface anything like what Excel has.
> It is true that XL's interface to fixed income analytics has very
> little real use.  However, to me that suggests that we make the XL
> functions a compatible as possible, and work in parallel on a more
> useful set of routines.
>  

May I propose the following:

For COUPNCD : when called with up to 4 arguments we try to get the same 
answers as XL, even when its answers are nonsense (which I believe they are)

ie
COUPNCD ("10/8/2001", "11/29/2002", 4, 0) == 11/28/2001
COUPNCD ("10/8/2001", "12/29/2002", 4, 0) == 12/29/2002

when called with a fifth argument (oem, specifying whether to hadle last 
of month maturity dates special) we can't be XL compatible anyways and 
we calculate the correct date, ie:

ie
COUPNCD ("10/8/2001", "11/29/2002", 4, 0, TRUE) == 11/29/2001
COUPNCD ("10/8/2001", "12/29/2002", 4, 0, TRUE) == 12/29/2002

and similarly for the other functions.


Andreas





--
Prof. Dr. Andreas J. Guelzow
http://www.math.concordia.ab.ca/aguelzow