https://bugs.documentfoundation.org/show_bug.cgi?id=55960
--- Comment #19 from b. <newbie...@gmx.de> ---
one more suggestion ... inspired from:
> 0.05*39=1.9500000000000002
> -2+0.05*39=-0.049999999999999822
the first calculation is 'two decimals' times 'integer' ...
users would expect, and mathematical correct would be! - a result with max two
decimal digits, thus just round to that,
same applies to second sample,
and e.g. 3,45 * 4,567 'two decimals' times 'three decimals' could / should
frankly be rounded to max six decimals,
addition / subtraction: round to max decimals of summands,
multiplication: round to product of count of decimals,
division: no chance,
exponentiation: simple exponents: round to exponent times decimals in mantissa,
fractional exponents: no chance,
radication: no chance,
other calculations ... are patched together from above?
handling like that won't be 'easy' - as the float's dont carry the info about
the source value they came from, but somewhere in the sheet you have the data
the user keyed in, at least in the file (or are values stored as floats there
too?),
and if you have that data calculation can be done by a function which keeps the
info about the 'original digits', and does appr. rounding on the result,
ok?
--
You are receiving this mail because:
You are the assignee for the bug.
_______________________________________________
Libreoffice-bugs mailing list
Libreoffice-bugs@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs
