Bug #62959 [Nab]: Wrong math
| From: | rasmus@php.net | Date: | Tue, 28 Aug 2012 18:19:10 +0000 |
| Subject: | Bug #62959 [Nab]: Wrong math | ||
| References: | 1 | Groups: | php.bugs |
| Request: | Send a blank email to php-bugs+get-172990@lists.php.net to get a copy of this message | ||
Edit report at https://bugs.php.net/bug.php?id=62959&edit=1
ID: 62959
Updated by: rasmus@php.net
Reported by: lauris dot kuznecovs at laurisnet dot com
Summary: Wrong math
Status: Not a bug
Type: Bug
Package: Math related
Operating System: Windows 7, Centos 5.8
PHP Version: 5.4.6
Block user comment: N
Private report: N
New Comment:
Depends on your definition of correct. Like I said, computers cannot represent
floating point with perfect accuracy. -2.8421709430404 E-14 is an extremely
small number very close to 0, so it is as close to "correct" as you are going to
get.
You should probably read through this http://floating-point-gui.de/ to get up to
speed on how computers treat floating point values. This isn't a PHP-specific
thing. For example:
Python 2.7.3 (default, Aug 1 2012, 05:14:39)
[GCC 4.6.3] on linux2
Type "help", "copyright", "credits" or "license" for more
information.
>>> (443.292 - 193.126 - 250.166)
-2.842170943040401e-14
Previous Comments:
------------------------------------------------------------------------
[2012-08-28 18:11:12] lauris dot kuznecovs at laurisnet dot com
Ok, lets say that the round() function works as expected, but still this
calculation (443.292 - 193.126 - 250.166) return: -2.8421709430404 E-14
And that is not correct!
------------------------------------------------------------------------
[2012-08-28 17:57:17] rasmus@php.net
-0 may not exist in traditional math, but it very much exists in computer
representations. Computers can't represent floating point values with complete
accuracy, so if you manipulate floating point you are going to get rounding
errors. -0 means a value rounded to 0 from the negative direction. You can read
http://en.wikipedia.org/wiki/Signed_zero for
more details.
------------------------------------------------------------------------
[2012-08-28 17:50:56] lauris dot kuznecovs at laurisnet dot com
Description:
------------
<? echo (443.292 - 193.126 - 250.166); ?>
Return: -2.8421709430404e-14
<? echo round(443.292 - 193.126 - 250.166, 2); ?>
Return: -0
But these calculations must return 0.
BTW -0 does not exist in MATH. Maybe some error in rounding function.
Test script:
---------------
<? echo (443.292 - 193.126 - 250.166); ?>
<? echo round(443.292 - 193.126 - 250.166, 2); ?>
Expected result:
----------------
0
------------------------------------------------------------------------
--
Edit this bug report at https://bugs.php.net/bug.php?id=62959&edit=1