https://bz.apache.org/ooo/show_bug.cgi?id=119374

--- Comment #10 from orcmid <[email protected]> ---
(In reply to Andreas Säger from comment #8)
> FYI: Recent discussion on the user forum about the meaning of negative time
> values which do make sense:
> https://forum.openoffice.org/en/forum/viewtopic.php?f=9&t=78985

I don't believe there is any concern, especially with respect to OpenFormula,
for cases that Excel does not handle.  However, the result for cases that Excel
does handle (all but the six negative results in the "sample test file") should
agree.

That is minimum interoperability.  ODF does not require implementation of
negative DayTime values, and the sample test file recognizes that.  

LibreOffice also matches Excel where Excel provides results, and uses an
extended DayTime range otherwise.  It seems ultimately pragmatic that AOO do
precisely the same thing as LibreOffice, providing interoperability among all
of the major ODF-supporting implementations and any software using the
historical result of Excel.

It is interesting that the handy identity T = INT(T) +
TIME(HOURS(T),MINUTES(T),SECONDS(T)) is preserved (isn't it?) whatever the
approach, since months and years are not considered, a good thing for many
reasons, and the HOURS, MINUTES, and SECONDS functions provide non-negative
results.

PS: The sample test file fails to obtain an usable result in Excel only where
the formula produces a negative result.  Where negative parameter values are
tested, and the result is non-negative, Excel provides the same result as the
formula. 

PPS: The OpenFormula specification explicitly allows the TIME implementation to
use the INT of each of its parameters.

PPPS: It would be useful to have interoperability notes that identify what
choices are made where the ODF OpenFormula specification is permissive.

-- 
You are receiving this mail because:
You are the assignee for the issue.

Reply via email to