Re: [office-comment] ODFF feedback

[Eike Rathke <erack@sun.com>]

Hi David,

old but not forgotten ;-)

On Wednesday, 2007-12-26 07:24:16 +0000, David King wrote:

> ==1==
> DATE has constraints:
> 1 <= Month <= 12; 1 <= Day <= 31
> 
> but the test cases show months rolling over - ie month = 13 is allowable 
> and not constrained to <=12.

Draft11 will have added the rollover behavior, which is common among
spreadsheet applications. Thanks.


> ==2==
> Each formula has test cases with headings Expression, Result, Level, 
> Comments. I cannot find where 'Level' is defined, so I think that this 
> is at least unclear. Very careful readings indicates that it probably 
> means the same as 'Group' - if so, why not use the same term?

It was discussed whether defining different conformance levels
additionally to the groups a function is in would allow applications to
implement a function up to some certain level, not necessarily having to
implement all aspects of a given function. It turned out this would be
overly complicated. We need to clean up the document, see also the TODO
paragraph of "2.3 Test Cases".


> ==3==
> ISLOGICAL
> Currently:
> "Semantics: If X is of type Logical, returns TRUE, else FALSE. 
> Applications that do not have a distinguished logical type will return 
> TRUE if ISNUMBER(X)."
> Better?:
> "Semantics: If X is of type Logical, returns TRUE, else FALSE. For 
> applications that do not have a distinguished logical type, ISNUMBER(X) 
> will return TRUE."

Yes, indeed. Will be in draft11.

> I'd also use the word 'distinct' rather than 'distinguished', but that 
> may be a difference between US and British English.

Sounds also better to me, not being a native English speaker at all.
There are also other places that talk of a distinct type, and others use
distinguished. Seems we need common language for this. I'll leave this
up to the native language editor ;-)

> It may also be that talking about ISNUMBER under ISLOGICAL is actually 
> unnecessary.

It is helpful, the expression AND(ISLOGICAL(TRUE());ISNUMBER(TRUE()))
can be used to determine whether the application has a distinct logical
type, the expression returns FALSE if it does, else TRUE.


> ==4==
> ISODD
> "Semantics: ...(per convert to number)."
> I've no idea what this means - may be a typo?

Should be "Conversion to Number" instead, section 6.2.4; same for
ISEVEN(). Will be in draft11.


> ==5==
> Excel will helpfully set a date display format if a cell contains eg 
> =TODAY(). Calc will helpfully display the result as a date, but without 
> apparently setting the format.
> 
> Does the ODFF standard need to address how similar formula results are 
> displayed? Another application might simply display the date/time serial 
> number.

The actual display of results may indeed vary between applications and
of course may be overriden by number styles as well. It is not the scope
of the formula specification to define application behavior on the
visual level.

  Eike

-- 
 OpenOffice.org / StarOffice Calc core developer and i18n transpositionizer.
 SunSign   0x87F8D412 : 2F58 5236 DB02 F335 8304  7D6C 65C9 F9B5 87F8 D412
 OpenOffice.org Engineering at Sun: http://blogs.sun.com/GullFOSS

PGP signature