Re: PI()

[Helmut Wollmersdorfer <helmut wollmersdorfer gmx at>]

Andreas J. Guelzow wrote:

> Seriously, I am wondering whether those XL rounding functions that don't
> behave sensibly could be packaged into their own little plugin, so that
> one could decide not to load them. In that case one wouldn't see them
> and keep looking to find INT and CEIL.

Looking into a detailed comparison between XL and Gnumeric, we can 
distinct the following situations:

1) XL and Gnumeric do the same - no problem
2) XL works wrong, and Gnumeric imitates the wrongness for compatibility
3) Gnumeric does more, e.g. is accepting more than a single cell as
   parameter, compared to XL which accepts only one cell. Compatibility
   XL -> Gnumeric works, but Gnumeric -> XL is a problem.
4) Gnumeric functions not existing in XL
5) XL is correct, but Gnumeric wrong - this is a bug

Apart from better documentation & testing, I have no genious idea for 
the remaining differences. Maybe a "XL compatibility option" will be a 
nice feature for those users which exchange files very often with XL-users.

Helmut Wollmersdorfer