Rob Weir, regarding 0^0:
> 1) If current ODF-supporting applications (OpenOffice, KSpread, Excel,
> Symphony, Gnumeric, Google Spreadsheets, etc.) agree on a value, then we
> should document it that way.
I don't think there's agreement. OpenOffice3, and traditional math, say
that 0^0 is 1. Excel 2003 says 0^0 is an error.
> 2) If there is no agreement in current spreadsheet usage, then indicate
> that it is "implementation-defined" or "undefined."
I think we can be narrower than that. Specifically, I think it can be
implementation-defined, but only permit a narrow set of options (e.g., 1 or Error).
--- David A. Wheeler
