Precise calculation in a string is possible if the user accepts a ratio 'n/d' instead of a rounded number. Since the input 'before' and 'after' are also floating point, they should also be presented as ratios. On the other hand, there is no added value in returning a string in this kata.

@Voile: what's your opinion on how to resolve this?

• Leaving strings aside and use approx equality?
• enforcing precise calculations? (how?)
• just drop the idea/kata?
• something else?

the inputs you give do not lead to the output you're talking about => ?

I forgot a multipler on days. For

before = 0.1
after = 0.2
days = 3 * 365

Some calculations gives 0.0375 while some gives 0.037500000000000006, which rounds to 0.037 and 0.038 respectively.

Moreover, if you use

before = 0.1
after = 0.6
days = 3 * 365

Then it's easier to get the exact result of 0.0875, which also rounds down to 0.087.

because that's a code working only for python 2

Definitely I could remove the string output - just thought it was maybe a bit of a way to flex some beginner f'strings in 7-8kyu at the same time but maybe just overcomplicates it. This is my first Kata so learning this as I go. Thanks for the feedback!

note: Actually, I think there is only one way to compute the correct result, in the case of your kata, so it's not about approximate equality.

@Voile: the inputs you give do not lead to the output you're talking about => ?

This code doesn't seem to pass when I copy and paste it.
range(1, 6) should equal [1, 2, 3, 4, 5]

?

the nromal course of action with floating point errors is to use approximate equality. But that suppose to have numbers as output (...that's where "you" realize that strings aren't a good output when you're actually interested in numbers)

=> requires to change the output type

Thanks. As all the random tests passed I thought it would be ok. What would you recommend - no rounding at all?

rounded to three decimal places

As usual, this is not recommended because of floating point errors. Example:

before = 0.1
after = 0.2
days = 365

This equals 0.0375 years, but rounds down to 0.037 due to floating point error.