digg

Facebook Connect Join Digg About

Google Calc fails: 399999999999999-399999999999998 = ?

google.com — They're doing it wrong.

Who dugg this?

Digg DudeWhat is Digg?

Digg is a place for people to discover and share content from anywhere on the web. Digg surfaces the best stuff as submitted and voted on by the community. Take the Tour to learn more.

35 Comments

show profanity settings
expand all
  • jjesusfreak01, on 09/25/2008, -0/+44Anyone consider that maybe Google is right, and you are wrong?
  • kaaresl, on 09/25/2008, -2/+33This is just a taste of what the LHC will unleash
  • Falldog, on 08/25/2008, -3/+28I found this one lazy afternoon. I didn't want to go outside and play baseball with the neighborhood gang and instead spent my time looking for minor application bugs. When 399999999999999-399999999999997 didn't work I tried 399999999999999-399999999999998. Heroism comes with patience. Dear Google, check please. *Colbert hand grab*
  • Osiris19, on 09/25/2008, -0/+21Guys, Google Calc sucks anyway, can't even divide by 0...

    Wait a minute...

    OH SHI-
  • enjoymoreradio, on 09/02/2008, -0/+19(X+1)-X returns 0 when X>333 333 333 333 333
  • Bilabrin, on 09/25/2008, -0/+11As a matter of fact, they are. I cross checked it on Wikipedia.
  • Wrathernaut, on 09/25/2008, -0/+11So apparently google runs on original pentium processors.
  • phorty40, on 09/25/2008, -0/+11they ARE billionaires after all.
  • kevyn, on 09/25/2008, -0/+9who actually found this out, and why were they running that calculation??
  • Abudman, on 09/23/2008, -0/+8333 333 333 333 335 - 333 333 333 333 334 = 0 is where it breaks.
  • jm1234567890, on 08/25/2008, -0/+8399999999999998-399999999999997

    also = 0
  • doulton, on 09/25/2008, -0/+5They finally beat that stupid 1=0 proof, or used it.
  • danielttt, on 09/25/2008, -0/+3Can you say Nobel Prize?...On behalf of myself and the rest of the free world, thank you Falldog...
  • Dracusis, on 09/26/2008, -0/+3I got it, but it wasn't even a little bit funny.
  • TheArssGuy, on 10/24/2008, -0/+3Digg people... do any of you have any idea how computers work?

    Calculators like Google Calc use floating point arithmetic to handle these calculations. In the case of Google Calc they're using double precision floating point arithmetic, which is why http://www.google.com/search?q=1.7976931348623157+ ... works but not http://www.google.com/search?q=1.79769313486232+*+ ...

    The reason why Google Calc cannot tell the difference between 333333333333335 and 333333333333334 is because at that point they have the same binary representation, because there's only that many significant digits the mantissa can represent. What's more puzzling however is that instead of having a standard mantissa of 52 bits as defined by the IEEE 754 standard it has 4 bits less, or 48 bits, whereas the exponent seems to have as many bits as the standard.

    So the real question is, 60-bit floating point representation WTF?

    EDIT : seems like Linux sometimes uses 60-bit "single precision" floats rather than 32-bit floats...
  • l33tforlif3, on 09/25/2008, -1/+3someone divided by zero...

    OH SH...
  • 87linux, on 10/12/2009, -0/+1Google is ALWAYS right. They are the future of all information, aren't they?
  • thejeremy, on 06/30/2009, -0/+1http://www.google-vs-bing.com/?q=399999999999999-3 ...

    !
  • Lazymoon, on 09/25/2008, -0/+1Evil does come in threes!

    (I know, I'll find the door)
  • deaper, on 05/15/2009, -0/+1399 999 999 999 990 - 39 999 999 999 989 = 3.6 × 10^14 WTF?
  • UCbizerkeley, on 09/25/2008, -0/+1bleh
  • laznik, on 09/25/2008, -1/+2this is a common problem with some languages, a famous example is PHP. It has to do with the way the language stores numbers internally as binary. The famous example is: floor((0.1+0.7)*10) will usually return 7 instead of the expected 8, since the internal representation will be something like 7.9
  • pajothadragon, on 09/25/2008, -1/+2shut up with ur desperate LHC comments for diggs loser
  • iamkid, on 09/26/2008, -0/+1well, most likely with it saying not 1, but 0, then its probably calculating large numbers like this in significant figures, just rounding the answer a little. or its just a glitch
  • Matt2k, on 09/25/2008, -1/+1I wouldn't call that common.

  • Quaidheizer, on 10/12/2009, -0/+01984 anyone?
  • ipekpromosyon, on 01/13/2009, -0/+0having a difference .. ipekpromosyon
  • lovespot2, on 06/12/2009, -0/+0is the last guy for real how old are you like 6 basic maths v v v basic maths
  • jcarlock, on 09/25/2008, -2/+1I found it starts having trouble after the 15th digit. Like its using a shorter variable type that necessary. Maybe 16 bytes, with one byte dedicated to sign and not changing to a string that converted back to an integer variable. I don't know. They'll fix it when it becomes an issue that people can't figure out this equation on there own.

    JC
  • rcoblentz, on 09/25/2008, -4/+2perhaps 'google' in the google calculator icon should be read as 'g00gle' (those are zeros, get it?!) haha, anyways...
  • egfx, on 09/25/2008, -4/+0here is a second taste http://www.ucubd.com/Index.aspx?id=730&cid=3516
  • Thok13, on 09/15/2008, -6/+1399 999 999 999 999 - 399 999 999 999 997 = 2 that works for me
  • DigitalCreators, on 08/24/2008, -9/+1small bug...FIX!
  • inactive, on 08/24/2008, -10/+1In your face!
  • SmoochiBoochie, on 08/24/2008, -13/+4YOUR SHIPMENT OF FAIL HAS ARRIVED.

  • Add a Comment

    Login or register to comment!
    Or, sign-in super fast with your Facebook account ...


© Digg Inc. 2009 — Content created and posted by Digg users is dedicated to the public domain | Terms of Use Updated | Privacy Policy
DIGG, DIGG IT, DUGG, DIGG THIS, Digg graphics, logos, designs, page headers, button icons, scripts, and other service names are the trademarks of Digg Inc.