FSHARP.md

Comparison with F#

This document compares this library with F#'s built-in unit support.

All F# code examples are from F# Units of Measure guide on MSDN.

Feature comparison

What F# has this library doesn't:

• full unit inference: (x,y,z) -> x*x + y + z*z*z is correctly inferred to have type (float<'u^3>, float<'u^6>, float<'u^2>) -> float<'u^6>

• more robust compile-time unit calculation, especially involving generics

• support for 32-bit floating point values

• cleaner syntax

What this library has F# doesn't:

• automatic unit conversion

• distinction between affine spaces and vector spaces

The main reasons F# unit support is better is that it's first-class language concept, not a library.

Syntax comparison

Defining units

F#:

[<Measure>] type cm
[<Measure>] type ml = cm^3

Scala:

type cm = DefineUnit[_c~:_m]
type ml = cube[cm]

Values

F#:

1.0<cm>
55.0<miles/hour>
55.0f<miles/hour> 

Scala:

1.0.of[cm]
55.0.of[mile/hour]
// 32-bit floating point numbers with units are not yet supported

Convertion functions

F#:

let convertg2kg (x : float<g>) = x / 1000.0<g/kg>

let length = 12.0<cm>
let x = float length

Scala:

def convertg2kg(x: DoubleU[g]): DoubleU[kg] = x / 1000.of[g/kg]
// explicitly stating the return type is not technically needed, but it makes Eclipse happy

// preferred way:
implicit val g_to_kg = one[kg].contains(1000)[g]

val length = 12.0.of[cm]
val x = length.value

Generics

F#:

let genericSumUnits ( x : float<'u>) (y: float<'u>) = x + y

type vector3D<[<Measure>] 'u> = { x : float<'u>; y : float<'u>; z : float<'u>}

let xvec : vector3D<m> = { x = 0.0<m>; y = 0.0<m>; z = 0.0<m> }

Scala:

def genericSumUnits[U: MUnit](x: DoubleU[U], y: DoubleU[U]) = x + y

case class Vector3D[U: MUnit](x: DoubleU[U], y: DoubleU[U], z: DoubleU[U])

val xvec = Vector3D[m](x=0.0.of, y=0.0.of, z=0.0.of)