Why 0.1 + 0.2 == 0.3 is false? | Hacker News Reader

Why 0.1 + 0.2 == 0.3 is false?

stackoverflow.com

15 pointswatermel0n9 comments

This headline is backwards -- it's the opposite of the actual stackoverflow question. The question is why, in the D language, `0.1 + 0.2 == 0.3` is true, when it should be false given the underlying floating point implementation.

(Spoiler alert: the answer is that the addition is done at compile time rather than run time.)

Douglas Crockford (JavaScript: The Good Parts) has talked about this problem with binary floating point to do decimal (human) math, on many occasions.

Here is one such instance. http://www.yuiblog.com/blog/2009/03/10/when-you-cant-count-o...

He has been on a crusade to introduce a decimal based floating point standard because he believes it is finally time we revisit this decision to use binary floating point.

Edit: spelling

On a side note, Excel is infuriatingly inconsistent: some operations work in a way that .1 + .2 == .3 (for example, try `=0.1+0.2=0.3`) but others fall apart (displayed as a fraction with one digit, 0.3 -> 2/7 but 0.1+0.2 -> 1/3)

floating point is hard. backwards compatibility is even harder

For a modern spreadsheet it would not be unreasonable for a user to expect symbolic arithmetic rather than floating point error soup.

In guile[1]:

  % guile
  guile> (= (+ 0.1 0.2) 0.3)
  #f

[1] - https://www.gnu.org/software/guile/

Guile(and most Lisps such as Common Lisp) support rationals specifically to avoid this problem, why not just use them ?

  scheme@(guile-user)> (= (+ 1/10 2/10) 3/10)
  $1 = #t

perl6 also use rationals (by default)...

  $ perl6

  > 0.1 + 0.2 == 0.3
  True

  > (0.1 + 0.2).perl
  3/10

For perl5 you must use the bigrat pragma:

  $ re.pl

  > 0.1 + 0.2 != 0.3
  1

  > use bigrat;
  > 0.1 + 0.2 == 0.3
  1

ref: http://perldoc.perl.org/bigrat.html

"why not just use them ?"

Sometimes decimal notation is the most natural, and sometimes you're copying numbers in decimal notation out of a book, wikipedia, or some other publication.

Reply on news.ycombinator.com