Re: [office-formula] A-functions

["David A. Wheeler" <dwheeler@dwheeler.com>]

Tomas Mecir wrote:
> On 9/7/06, David A. Wheeler <dwheeler@dwheeler.com> wrote:
>> I think you're forgetting that logical types are distinct types in
>> some apps (like Excel, Gnumeric, and SheetToGo).
>> Yet _ALL_ applications handle this identically:
>>  SUM([.A1:.B3]; [.C4]<10)
>> because the second number is a logical, and inline logicals ARE 
>> converted
>> to numbers, even when logical values are distinct types.
>> This construct can be used to include compensating factors.
>
> Wouldn't it then be sufficient to follow the way of OOo, and make the
> non-A versions always include logical values as well ? Both reference
> and inline ?
No, because we have ALREADY agreed
that whether or not logical types are distinct types is
implementation-defined - after MANY months of discussion.
The semantics as written for SUM, etc.,
are one of the consequences of that decision.
If we changed to the kind of semantic you suggest, then many apps
including Gnumeric, Excel, SheetToGo, and many others would probably
NEVER comply with the spec, either now OR in the future.

I'm not really eager to revisit that discussion, once was painful enough 
:-).

We can easily add specific functions with different semantics.
But I don't think it's reasonable to demand particular semantics for
nearly all functions when we KNOW that many implementations implement
CONFLICTING semantics, will probably ALWAYS conflict with them,
and what's worse, that there are almost certainly many documents that
DEPEND on these DIFFERENT semantics.

--- David A. Wheeler