2009 Honda Fit: Time for Maths Camp

omg basic calculator failz lololol

looks like a basic programming error. Certain programming languages have seemingly bizarre ways of storing numbers, so the outputs have to be specially formatted to avoid issues like this.

using the numbers in the photos, i could argue that the calculator is correct. the question is: Is 0.9999999999999999... = 1? the answer, is yes. we all learned that 1 ninth (1/9) is 0.111111... and 2/9 is .2222222... and so on. thus nine ninths (9/9) would be 0.99999999... but we also learned that any time the numerator and the denominator are the same that it is equal to 1. so, math nerds, the calculator fails to distinguish these fourth grade math lessons. just round up.

Floating point arithmetic in computers is not an exact science. The way that floating point numbers are represented internally in computers is inexact, at best. As a result, the math performed on these numbers is also inexact. Everything dealing with floating point in a computer is an approximation. Thus the puzzling answer given by the calculator.

I'd give more computer science-y details, but I'm afraid that I've already put everyone to sleep. Maybe I should have thrown a Megan Fox analogy in there somewhere. :-]

Should be 33.29. You got it wrong too. ;)

jefe2000 -
I think the analogy your looking for has to do with the junk left if you don't trunk your variable.

@ jefe2000 beat me to it.

MS Excel does crap like this all of time

If this is the same as the one in some Acuras (as it looks to be), I always liked the fuel economy converter, just for giggles to see my consumption in liters per 100 km

80085

As jefe2000 said, this is a classic example of floating point math in computers. Much of the problem comes from the fact that it is impossible to represent 0.1 in binary (or, in fact, any fraction p/q where q is not a power of 2).

Thus, the usual example of comparing two reals as you do here, none of the three numbers can be represented exactly in binary, so, quite often, you get the oddity that x+y doesn't equal z=(x+y). This is also why you need to get used to comparing reals not by equality but by seeing if their difference is less than a threshold: |x+y-(z)| < 1e-6.

You also find the associative {(a+b)+c = a+(b+c)} and distributive {(a+b)*c = a*c+b*c} properties don't hold true with floating point math.

Maybe this explains why the Fit's mpg meter is off by about 10%.

OMG NERD ALERT!!!
Your friendly neighbor who has two engineering degrees

jefe2000 and fortran are correct. If you'd like to learn more about it you can read this article:
http://www.techradar.com/news/computing/why-computers-suck-at-maths-644771

fortran nailed it. Anything that cannot be exactly represented as a summation of powers of 2 is impossible to represent on a computer.

doesn't matter how the computer internally represents it.
By the time it hits the display screen to interact with the human, it should have hit a round() or trunc() and displayed it right.

$1 Made in China calculator can do it.
Still a goof regardless of what the computer's internal computing does.

This programmer massive failz Intro to Computer Science 101 first week project.
Like omg he needs additionaly tutoring.
Go to the prof's office and seek help. Ask for 1 week deadline extension.

Despite Fortran's explanation, I'm not convinced. I've never heard of a calculator that had this error. It's my understanding that calculators convert from binary to decimal prior to calculations being performed.

Is the calculation error repeatable?

Oh come on, Vermont ain't that bad! You have to shoo away the hippies with a broom but otherwise it's a nice state.

My Tribeca has a calculator and it can convert PSI to Kilopascals; not that you would ever need that. It also has F to C, and others but it does round unlike the fit.

estreka: fortran's explanation is completely correct. Here's an example of a calculator that has this error: http://www.amazon.com/review/R39056TFUDL65M/ref=cm_cr_pr_viewpnt#R39056TFUDL65M

There's no real way for a calculator to "convert from binary to decimal."

Yeah but it doesn't change the fact using floating-point arithmetic in a calculator application is pretty lazy. Imagine if your bank used a calculator like that.