[odf-discuss] Formula for Disaster

[Daniel Carrera]

Andreas J. Guelzow wrote:
> Interesting but unfortunately wrong.
> 
> The definition of CEILING in MS Excel or MSOOXML can not be
> mathematically wrong. It may be different from the definition of ceiling
> used in Mathematics but being a definition this is simply an agreement
> for the moment.

I know what you mean, but to take a more extreme case, what if ECMA says 
"The spreadsheet function SIN() is defined to return 12 whenever its 
argument is greater than 1 and 0 otherwise". You could still say that 
this is not mathematically wrong because they are just defining a 
function and it just happens to be different from the definition of the 
mathematical sin function.

The point is that I do think we should expect spreadsheet function 
definitions to match (within reason) the mathematical functions of the 
same name.

I don't have a rigorous definition of "within reason". I'm thinking of 
things like the fact that the Openformula CEILING function includes an 
additional parameter, but it is nonetheless true that:

CELING(x ; 1) = CEIL(x)

Where 'CEIL' is the mathematical ceiling function and 'CEILING' is the 
OpenFormula CEILING function. This is something I'd consider a 
reasonable extension of the mathematical ceiling function.

If there is no 'n' value such that CEILING(x ; n) = CEIL(x) for all x I 
would consider that a problem. And this is the case in Ecma/MSXML.

Cheers,
Daniel.