Re: [office-formula] Matrix evaluation

[Patrick Durusau <patrick@durusau.net>]

Eike,

Thanks!

A couple of more questions below:

Eike Rathke wrote:
> Hi Patrick,
>
> On Sunday, 2010-01-24 09:55:15 -0500, Patrick Durusau wrote:
>
>   
>>> If the expression is being evaluated in a cell flagged as a being part  
>>> of a 'Matrix' (ODF 8.1.3 |table:number-matrix-columns-spanned|):
>>>
>>>       
>> under 2) Matrix Evaluation
>>
>> is correct but it isn't a requirement for matrix evaluation.
>>
>> That is to say that OpenFormula is defining matrix evaluation for all  
>> instances of matrix evaluation and not just those in a cell flagged....
>>
>> Yes?
>>     
>
> Yes. Matrix/array evaluation also occurs if the expression is part of
> a ForceArray parameter.
>
> I propose to change that to
>
> "If the expression is evaluated in array context:"
>
> and define when an array context is given, i.e. the cell is flagged
> being part of a Matrix or the expression is part of a ForceArray
> parameter.
>
>   
>> Which makes the relationship between 2.2, which says it is a  
>> generalization of 2.1 even odder.
>>     
>
> I don't think so. 2.2 talks about the behavior when a function does not
> expect array parameters but evaluation occurs in array mode. The rules
> of 2.1 are applied to obtain the elements for each iteration. This is
> generalization in the sense of applying the specific matrix/array rules
> to functions that do not take array parameters.
>
>   
OK, but 2.1 talks about rules for storing non-scalar results.

2.2 describes a processing model to be used when non-scalar inputs are 
passed to scalar functions. Such as describing a result matrix and its 
size. 2.2.2 (I would delete the "unexpected" language as unnecessary.)

I understand that it is necessary to describe the storage of the results 
of non-scalar inputs to scalar functions but I think it would be clearer 
to state:

1) Rules for storing non-scalar results (and clearing up the use of 
source, etc.)

2) Rules for evaluation of non-scalar inputs by scalar functions, 
including the storage of the results of such evaluations.

Yes, there are similarities and it could be argued that one is a 
generalization of the other but in the interest of clarity I would 
suggest that we separate them here.

Hope you are having a great evening!

Patrick






>> Was it intended to have one matrix evaluation for ODF and another more  
>> generally?
>>     
>
> No.
>
>   
>> I suspect I need a conference call with the author of this section to  
>> sort this out.
>>     
>
> He isn't a member of the committee anymore, but I think I can answer
> your questions..
>
>   Eike
>
>   

-- 
Patrick Durusau
patrick@durusau.net
Chair, V1 - US TAG to JTC 1/SC 34
Convener, JTC 1/SC 34/WG 3 (Topic Maps)
Editor, OpenDocument Format TC (OASIS), Project Editor ISO/IEC 26300
Co-Editor, ISO/IEC 13250-1, 13250-5 (Topic Maps)