Re: [office-formula] 0^0 in OpenFormula and OOXML

[Andreas J Guelzow <aguelzow@math.concordia.ab.ca>]

On Sun, 2009-01-25 at 15:11 -0500, Patrick Durusau wrote:
> Greetings!
> 
> While on the road I have started compiling a rough comparison of 
> OpenFormula and OOXML operator and functions definitions.
> 
> It didn't take long to reach "^" and we report in the notes in 
> OpenFormula that Excel returns #NUM!,

for 0^0.

>  which may well still be true, I 
> don't have it on my laptop, but I was interested to find that OOXML says 
> in 18.17.7.113 EXP, in its examples that EXP (0) results in 1.

EXP(0) has absolutely nothing to do with 0^0. Surely EXP(0) is 1 in any
self-respecting spreadsheet.

> 
> And: "However, if x is too large for the result to be representable, 
> #NUM! is returned."
> 
> So, I would not think that "0^0" would violate that rule.

??? I am really not sure what you are saying!?

Andreas

> 
> I did find 18.17.7.255 POWER ( x, y ) which does say that if:
> 
> "x is zero and y is less than or equal to zero, #DIV/0! is returned."
> 
> (Note the ambiguous use of "whole number" in the preceding list item: 
> "The value of x is negative and y is not a whole number, #NUM! is 
> returned." Hard to say if that includes or excludes 0.)
> 
> Even though our note will be hidden, its best not to have 
> misapprehensions about what others have in fact said.
> 
> In this particular case, it appears to me that OOXML has:
> 
> Failed to define the result for the case of 0^0 for the exponent 
> operator ^ other than perhaps by implication from 18.17.7.255 POWER but 
> then that would contradict the definition of ^ (18.17.7.113 EXP) which 
> clearly allows 0 as both components. (It wasn't just the example that 
> was the tip off. If you run the regex to ground you find that 0 is 
> accepted in both positions.)
> 
> On the other hand, OpenFormula has failed to account for the result 
> "#DIV/0!" That is assuming OpenFormula should be recording the current 
> results of spreadsheet behavior as a "standard" in a rather loose sense 
> of the word.
> 
> And, note that OpenFormula under 6.15.45 Power, fails to define the 
> result in the case of POWER ( 0; 0).
> 
> Interesting that the two standards fail to make the same definition but 
> at opposite locations and leave it only implied to implementers, 
> assuming they have read closely enough to realize there is an issue.
> 
> Hope everyone is having a great day!
> 
> Patrick
> 
> PS: It seems to me that we need to answer the question of if and where 
> OpenFormula differs from OOXML and why? We can either answer that now or 
> we can wait for it to come up upon submission of ODF 1.2 to ISO. From an 
> editor's point of view I would prefer now.

-- 
Andreas J. Guelzow
Concordia University College of Alberta