Hi, quick question about doubles. In Kotlin, `0.1 ...

# getting-started

[Tony Blundell]

Hi, quick question about doubles. In Kotlin, 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 < 0.8 evaluates as true, because the left-hand side is actually 0.79r instead of 0.8. What's the best way of dealing with this? I guess I could convert to a string and back, but that feels a little 'hacky'. Is there a better way?

[zpearce]

If you need accuracy, BigDecimal is the way to go

[ephemient]

https://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.html

[ephemient]

if you're going to stick with floating point, all comparisons should have an error bound

[ephemient]

also it's not quite 0.79r or 0.8; neither is exactly representable in floating point

[Tony Blundell]

👍 thanks both

[Denis Korotchenko]

You can also try to use construction like this:
if (abs(res-pred) < eps)...
where eps is a constant like 10^(-7) etc.

[elizarov]

Highly recommended read: https://floating-point-gui.de/