[embedded_hiker]

This put my daughter off of programming. When she was 7, I showed her how to use python in immediate mode, and she got it without difficulty. She even understood variables. Then one day she wanted to add prices, and she got one of these errors, and she never wanted anything to do with it again.

[angry_moose]

I still remember writing something in high school along the lines of:

  i=0  
  while(i<1):  
      <something with i>  
      i=i+.1

And spending hours trying to figure out why it ran an extra iteration, and this was early enough it wasn't easily googleable. Whatever I was doing with i needed it to be .1, .2, .3... and thought I was being clever not doing 1...10 and dividing by 10 every iteration within the loop. I think there was also a weird language quirk with whatever I was using that a print(i) rounded to a handful of decimal places so it looked fine while debugging it.

Very frustrating, but in retrospect very eye opening.

[Lendal]

Okay, I'll defend floating point numbers. The choice of floating point over decimal represents the choice of science over money. In science, it's more important to have a number system that represents everything from the infinitely small to the infinitely large, rather than one that has perfect precision. Because in nature, perfect precision does not exist. It doesn't matter what pi is to perfect accuracy because there are no perfectly round circles in reality. Only in money and mathematics do people really care about perfect precision. In the real world, precision is negotiable.

I think that's a good lesson for kids.

[rhn_mk1]

Yes, but actually no. Precision becomes important once you start digging in. Calculate the GPS time dilation without sufficient precision and you'll be in trouble. Go down to quantum physics to discover that the exact ratio of mass between the electron and proton might matter for your nuclear reactor.

[worksonmine]

You don't even need to get that specific, even web developers encounter this sooner or later, often as a UI bug in what should be really simple math. Then one day you wonder "what the fuck are all these zeroes? Oooooh..."

That's how I learned about it years ago.

[londons_explore]

It's understandable - you trust a tool like a calculator to give you the right answer. If it sometimes makes mistakes and you have to check each answer by hand, it isn't really saving you any time.

To many, a rounding error makes the answer "wrong", and suddenly the tool has switched from a reliable one into an untrustworthy one.

[dahfizz]

> you trust a tool like a calculator to give you the right answer.

By middle school, kids should have learned that you can't trust calculators. There are all sorts of numbers like pi, e, sqrt(2) that are impossible to represent. Once you start getting into trig, you have to accept rounding.

[aaaronic]

Sure, but .1 is definitely representable, so they can be excused for finding it a little unreasonable that .1+.1+.1+.1+.1+.1+.1+.1+.1+.1 doesn't equal 1 in many languages.

Explaining _why_ .1 isn't representable requires explaining IEEE-754 and explaining _that_ requires an understanding of binary numeric representation.

I teach college students who find this confusing, so I think it's fair that the average person finds floating point behavior confusing (in fact, I've had to explain to Physics Professors doing computation simulation work why their 1-<tiny number> isn't working out the way they expect -- though they initially tried using double doubles to get around the problem).

[chrchang523]

This does depend a bit on the calculator. embedded_hiker's anecdote has made me update in the direction of exposing my daughter to Wolfram Alpha before Python...

[jdmichal]

My AP calc teacher was an expert on writing tests that would trigger calculators into approximation mode. Pretty much every homework problem, the calculator could easily do an exact answer. But you better have learned, because come test time, the best your calculator is going to offer is 0.942858934759084...