Re: [office-formula] EFFECT and NOMINAL

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

Eike:
> OOo has EFFECTIVE and EFFECT_ADD vs. Excel EFFECT, and NOMINAL and
> NOMINAL_ADD vs. Excel NOMINAL, where the _ADD functions are compatible
> with the Excel functions and compared to the OOo internal functions
> differ only in the display format of the returned value, the OOo
> internal functions display it as percentage whereas the Excel functions
> don't. OOo probably can change implementation to use the Excel
> compatible functions and we don't need to distinguish.

Great!   Merging into a single definition for EFFECTIVE and NOMINAL is absolutely DELIGHTFUL.

Indeed, from a formula point-of-view that's out of scope.  The default display format isn't really specified - just the calculated result.  Indeed, different versions of the same app (such as Excel) change what the default format is for a newly-defined cell; those tend to be heuristics, and don't matter when exchanging documents (it's purely a user-interface issue). 


> However, NOMINAL currently is in the Medium Group and EFFECT in the
> Large Group. Shouldn't both be in the same group? Also, NOMINAL test
> cases say level 3 while EFFECT test cases say level 1. So how do we
> define it correctly?

The starting place for the "Medium" group was the set of functions implemented by at least 4 of the following 5 applications: Excel 2003, Gnumeric, Lotus 1-2-3 v9, OpenOffice.org 2, and (WordPerfect) QuattroPro 12.  Think of this as a voting algorithm; if so many separate implementors implement the function, then it's likely to be important for interoperability.  We then modify the list as the group deems appropriate.

Under this notion, NOMINAL is in the list, and EFFECT is not.  But I'm happy adding EFFECT to the medium group, it makes sense.

So I'm going to define NOMINAL and EFFECT, place them both in medium, and adjust the notes so it's mentioned that NOMINAL_ADD and EFFECT_ADD of OOo are old versions.  Any objections?

--- David A. Wheeler