```var x: Number = 2 var y: Number = 3 var c = x ...

# getting-started

[Ray Rahke]

var x: Number = 2
var y: Number = 3

var c = x + y

it is so annoying that this is not accepted in kotlin

[Joffrey]

Why do you use the Number type?

[Ray Rahke]

@Joffrey because the vast majority of the time that I am annotating a function parameter or a class field as something numerical, it is indifferent to whether it is a float or integer

[Ray Rahke]

most high level programming does not care about this distinction

[Ray Rahke]

i want to be able to do fn(5) bu also fn(5.9)

[Ray Rahke]

and kotlin does not accept x: Int | Double union types

[Ray Rahke]

I'll be interfacing with some game framework and want to type player.health -= 5

[Ray Rahke]

and then it goes "woops! you need to use a double!"

[Ray Rahke]

so I have to go
player.health -= 5.0

[Ray Rahke]

and that is so pedantic.

[Joffrey]

If you don't care about the difference, why not use a float or double in the first place?

[Ray Rahke]

if I use a double, then that forces all of the consumers of my api to pass in a double later

[Ray Rahke]

imagine I have some function that returns an integer

[Ray Rahke]

and for whatever resaon I want to pass the result of that function to some gameplay logic that moves a player

[Ray Rahke]

but I annotated that player class as using doubles

[Ray Rahke]

now there is this stupid conversion step I need to do

[Ray Rahke]

i end up having to make literally all of my codebase default to doubles

[Ray Rahke]

just in case i might want to interface between a module that uses doubles

[Joffrey]

and then it goes "woops! you need to use a double!"

It doesn't. You can totally do this.
https://pl.kotl.in/nnmxU1HGP

[Ray Rahke]

oh I was wrong about that

[Ray Rahke]

i see. it must have been when I had this

var x: Double
var y: Number = 5

x + y

[Ray Rahke]

you cant add a Number to a double

[Joffrey]

What would this operation return? In general you should never use Number for your types, if your function manipulates numbers that can be non-integers, just use double or float, so the operations you do are clearly defined in terms of overflow and return type. If you're fine with languages like JS that don't distinguish different number types, then you pretty much get doubles everywhere, so the argument about ints is moot. In general though, we want to be aware of the precision or overflows of the operations we perform. So we use distinct types and explicit conversions where it makes sense

[Arjan van Wieringen]

But most dynamic language actually DO care about that distinction. But they bury it under implied assumptions leading to a world of pain.

[Landry Norris]

You actually can also create an extension operator function for Number.plus(Number) that will explicitly define how the different types interact using a when expression.

[Shawn]

relevant thread https://kotlinlang.slack.com/archives/C0B8Q383C/p1641757248000700

[CLOVIS]

the vast majority of the time that I am annotating a function parameter or a class field as something numerical, it is indifferent to whether it is a float or integer

As long as you have that attitude, please don't work in finance or any other critical fields. And read a bit on floating point arithmetic.