Why don’t my floating-point numbers add up?

Recently happened upon this friendly site explaining the issues regarding floating-point math on binary computers and thought it useful.  As I'm sure you're aware it takes special talent to take a complicated subject and present it in an easy to understand fashion.

The article may be found below:

• What Every Programmer Should Know About Floating-Point Arithmetic

Since there is no page specifically for Python at the site, I thought I'd whip up this complement to the piece.  Below is an example for the language.  Note, that first half illustrates standard floating-point math and its associated issues.  The second half uses the decimal module to alleviate them. The drawback, as always a cost in performance.

1 2 3 4 5 6 7 8 9 10 11

Python 2.6.2 ... >>> >>> .1 + .2 0.30000000000000004 >>> .33*3 0.98999999999999999 >>> from decimal import Decimal >>> >>> Decimal('.1') + Decimal('.2') Decimal('0.3')

Further Reading:

• The most commonly forwarded link on the subject is Sun's much more detailed, though accessible piece.  Study this one and earn your black belt (1st Dan):
  What Every Computer Scientist Should Know About Floating-Point Arithmetic