[Libreoffice-bugs] [Bug 128312] Calculation involing some decimals return incorrect floats with 12 d.p.

bugzilla-daemon at bugs.documentfoundation.org bugzilla-daemon at bugs.documentfoundation.org
Tue Oct 22 06:37:46 UTC 2019


https://bugs.documentfoundation.org/show_bug.cgi?id=128312

Sam <sergeroz at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
     Ever confirmed|0                           |1
         Resolution|NOTABUG                     |---

--- Comment #3 from Sam <sergeroz at gmail.com> ---
(In reply to Mike Kaganski from comment #2)
> This is normal: computers represent numbers as binary numbers, and not all
> decimal numbers are representable as finite binary floating-point numbers.
> E.g., a simple 0.1 is not finite in binary form.
> 
> See http://en.wikipedia.org/wiki/Floating_point#Accuracy_problems.
> 
> Users working with fractional numbers should account for that.
> 
> Closing NOTABUG.

We can't seriously expect home & business users to know about IEEE 754,
'floating point' and 'mantissa'. A significant number of people have very
limited knowledge of math. What our users will notice is that 31000.99 -
32000.12 does not equal -999.129999999997.

Microsoft Excel and Google Sheets do not exhibit this behaviour. Even VisiCalc
from 1981 (https://www.pcjs.org/apps/pcx86/1981/visicalc/) does not exhibit
this behaviour.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/libreoffice-bugs/attachments/20191022/ed012832/attachment-0001.html>


More information about the Libreoffice-bugs mailing list