Agent Skills: Convert Elm to F#

Convert Elm to F#

Convert Elm code to idiomatic F#. This skill extends meta-convert-dev with Elm-to-F# specific type mappings, idiom translations, and tooling for migrating functional frontend code to the .NET ecosystem.

This Skill Extends

• meta-convert-dev - Foundational conversion patterns (APTV workflow, testing strategies)

For general concepts like the Analyze → Plan → Transform → Validate workflow, testing strategies, and common pitfalls, see the meta-skill first.

This Skill Adds

• Type mappings: Elm types → F# types
• Idiom translations: Elm patterns → idiomatic F#
• Error handling: Elm Result/Maybe → F# Result/Option
• Architecture: The Elm Architecture (TEA) → F# Elmish or MVU patterns
• Functional patterns: Pure functions, immutability preserved in F#

This Skill Does NOT Cover

• General conversion methodology - see meta-convert-dev
• Elm language fundamentals - see lang-elm-dev
• F# language fundamentals - see lang-fsharp-dev
• F# web frameworks in depth - see framework-specific skills

Quick Reference

| Elm | F# | Notes | |-----|-----|-------| | String | string | Direct mapping | | Int | int | 32-bit signed integer | | Float | float | 64-bit floating point | | Bool | bool | Direct mapping | | List a | 'a list | Immutable linked list | | Array a | 'a array | Mutable fixed-size array | | Maybe a | 'a option | Same semantics: Some/None | | Result error value | Result<'value, 'error> | Same semantics: Ok/Error | | type alias | type (type alias) | Direct mapping | | type (union) | type (discriminated union) | Direct mapping | | Cmd msg | Async<'msg> or Elmish Cmd<'msg> | Depends on framework | | Sub msg | Event subscriptions in Elmish | Framework-specific | | Html msg | Elmish.React ReactElement | Frontend framework |

When Converting Code

1. Analyze source thoroughly before writing target
2. Map types first - Elm and F# type systems are very similar
3. Preserve functional purity - both languages emphasize immutability
4. Adopt F# idioms - leverage .NET libraries and computation expressions
5. Handle The Elm Architecture - map to Elmish or custom MVU implementation
6. Test equivalence - same inputs → same outputs

Type System Mapping

Primitive Types

| Elm | F# | Notes | |-----|-----|-------| | String | string | UTF-16 in F# (via .NET), UTF-8 in Elm | | Int | int | Both 32-bit signed | | Float | float | F# float is 64-bit (alias for double) | | Bool | bool | Direct mapping | | Char | char | Single character | | () | unit | Unit type (void) | | never | N/A | Elm's impossible type; F# doesn't have exact equivalent |

Collection Types

| Elm | F# | Notes | |-----|-----|-------| | List a | 'a list | Immutable linked list, same semantics | | Array a | 'a array | F# arrays are mutable but similar performance | | Set a | Set<'a> | Immutable set in both | | Dict k v | Map<'k, 'v> | Immutable map; F# also has Dictionary<'k,'v> (mutable) | | Tuple | Tuple | Same syntax: (a, b) |

Composite Types

| Elm | F# | Notes | |-----|-----|-------| | type alias | type (type alias) | type Person = { Name: string; Age: int } | | type (union) | type (discriminated union) | Same concept, slightly different syntax | | Record | Record | Same semantics, nearly identical syntax | | Opaque type | Single-case union | type Email = Email of string |

Option/Maybe Types

| Elm | F# | Notes | |-----|-----|-------| | Maybe a | 'a option | Both have Some/Just and None/Nothing | | Just x | Some x | Constructor name differs | | Nothing | None | Constructor name differs | | Maybe.withDefault | Option.defaultValue | Same semantics | | Maybe.map | Option.map | Same semantics | | Maybe.andThen | Option.bind | Monadic bind |

Result Types

| Elm | F# | Notes | |-----|-----|-------| | Result error value | Result<'value, 'error> | Type parameters in reverse order! | | Ok value | Ok value | Same constructor | | Err error | Error error | F# uses Error, not Err | | Result.map | Result.map | Same semantics | | Result.andThen | Result.bind | Monadic bind | | Result.withDefault | Result.defaultValue | Same semantics |

Critical Note: Elm's Result error value has error first, but F# Result<'T, 'TError> has value first!

Generic Type Parameters

| Elm | F# | Notes | |-----|-----|-------| | a, b, c | 'a, 'b, 'c | F# uses single quote prefix | | comparable | 'a when 'a : comparison | Constraint syntax differs | | number | ^a when ^a : (static member (+) : ^a * ^a -> ^a) | F# uses SRTPs (complex) | | appendable | No built-in equivalent | Manual trait constraints |

Idiom Translation

Pattern 1: Maybe/Option Handling

Elm:

type alias User =
    { name : String
    , email : Maybe String
    }

getUserEmail : User -> String
getUserEmail user =
    Maybe.withDefault "No email" user.email

findUser : Int -> List User -> Maybe User
findUser id users =
    List.filter (\u -> u.id == id) users
        |> List.head

F#:

type User = {
    Name: string
    Email: string option
}

let getUserEmail (user: User) : string =
    user.Email |> Option.defaultValue "No email"

let findUser (id: int) (users: User list) : User option =
    users |> List.tryFind (fun u -> u.Id = id)

Why this translation:

• Maybe → option is a direct semantic mapping
• Just → Some, Nothing → None
• Maybe.withDefault → Option.defaultValue
• List.head : List a -> Maybe a → List.tryFind or List.tryHead
• F# has richer Option module with more combinators

Pattern 2: Result/Error Handling

Elm:

type alias Error = String

parseAge : String -> Result Error Int
parseAge str =
    case String.toInt str of
        Just age ->
            if age >= 0 then
                Ok age
            else
                Err "Age must be non-negative"
        Nothing ->
            Err "Not a valid integer"

validateUser : String -> String -> Result Error User
validateUser name ageStr =
    parseAge ageStr
        |> Result.andThen (\age -> Ok { name = name, age = age })

F#:

type Error = string

let parseAge (str: string) : Result<int, Error> =
    match System.Int32.TryParse(str) with
    | true, age when age >= 0 ->
        Ok age
    | true, _ ->
        Error "Age must be non-negative"
    | false, _ ->
        Error "Not a valid integer"

let validateUser (name: string) (ageStr: string) : Result<User, Error> =
    parseAge ageStr
    |> Result.bind (fun age -> Ok { Name = name; Age = age })

Why this translation:

• Result error value (Elm) → Result<'value, 'error> (F#) - note reversed type parameters
• Ok/Err → Ok/Error
• Result.andThen → Result.bind
• F# uses TryParse pattern for parsing (returns tuple bool * value)
• Both support railway-oriented programming with bind/map

Pattern 3: Union Types and Pattern Matching

Elm:

type Msg
    = Increment
    | Decrement
    | SetValue Int
    | Reset

update : Msg -> Model -> Model
update msg model =
    case msg of
        Increment ->
            { model | count = model.count + 1 }

        Decrement ->
            { model | count = model.count - 1 }

        SetValue n ->
            { model | count = n }

        Reset ->
            { model | count = 0 }

F#:

type Msg =
    | Increment
    | Decrement
    | SetValue of int
    | Reset

let update (msg: Msg) (model: Model) : Model =
    match msg with
    | Increment ->
        { model with Count = model.Count + 1 }

    | Decrement ->
        { model with Count = model.Count - 1 }

    | SetValue n ->
        { model with Count = n }

    | Reset ->
        { model with Count = 0 }

Why this translation:

• Discriminated unions are nearly identical
• Elm: Type arg → F#: Type of arg
• Record update syntax: { model | field = value } → { model with Field = value }
• Pattern matching syntax is nearly identical
• F# requires explicit of keyword for union cases with data

Pattern 4: List Operations

Elm:

users : List User
users =
    [ { name = "Alice", age = 30 }
    , { name = "Bob", age = 25 }
    ]

activeUsers : List User -> List User
activeUsers =
    List.filter .active
        >> List.map .name
        >> List.sort

totalAge : List User -> Int
totalAge users =
    List.foldl (\user sum -> sum + user.age) 0 users

F#:

let users: User list =
    [ { Name = "Alice"; Age = 30 }
      { Name = "Bob"; Age = 25 } ]

let activeUsers: User list -> string list =
    List.filter (fun u -> u.Active)
    >> List.map (fun u -> u.Name)
    >> List.sort

let totalAge (users: User list) : int =
    users |> List.fold (fun sum user -> sum + user.Age) 0

Why this translation:

• List syntax nearly identical: [ items ]
• Function composition: >> in both languages
• List.foldl → List.fold (F# fold is left-associative by default)
• F# has more list functions (List.sumBy, List.groupBy, etc.)
• Record access: .field → (fun r -> r.Field) or use lambda

Pattern 5: The Elm Architecture (TEA) → F# Elmish

Elm:

type alias Model =
    { count : Int }

type Msg
    = Increment
    | Decrement

init : ( Model, Cmd Msg )
init =
    ( { count = 0 }, Cmd.none )

update : Msg -> Model -> ( Model, Cmd Msg )
update msg model =
    case msg of
        Increment ->
            ( { model | count = model.count + 1 }, Cmd.none )

        Decrement ->
            ( { model | count = model.count - 1 }, Cmd.none )

view : Model -> Html Msg
view model =
    div []
        [ button [ onClick Decrement ] [ text "-" ]
        , text (String.fromInt model.count)
        , button [ onClick Increment ] [ text "+" ]
        ]

F# (using Elmish):

open Elmish

type Model = { Count: int }

type Msg =
    | Increment
    | Decrement

let init () : Model * Cmd<Msg> =
    { Count = 0 }, Cmd.none

let update (msg: Msg) (model: Model) : Model * Cmd<Msg> =
    match msg with
    | Increment ->
        { model with Count = model.Count + 1 }, Cmd.none

    | Decrement ->
        { model with Count = model.Count - 1 }, Cmd.none

let view (model: Model) (dispatch: Msg -> unit) =
    div [] [
        button [ OnClick (fun _ -> dispatch Decrement) ] [ str "-" ]
        str (string model.Count)
        button [ OnClick (fun _ -> dispatch Increment) ] [ str "+" ]
    ]

Why this translation:

• Elmish is F#'s implementation of The Elm Architecture
• Model, Msg, init, update, view pattern is identical
• Cmd type is similar but uses F# async under the hood
• View function receives dispatch explicitly in F#
• Html msg → ReactElement (via Fable.React)

Pattern 6: JSON Decoding

Elm:

import Json.Decode as Decode

type alias User =
    { id : Int
    , name : String
    , email : Maybe String
    }

userDecoder : Decode.Decoder User
userDecoder =
    Decode.map3 User
        (Decode.field "id" Decode.int)
        (Decode.field "name" Decode.string)
        (Decode.maybe (Decode.field "email" Decode.string))

F#:

open System.Text.Json
open System.Text.Json.Serialization

type User = {
    [<JsonPropertyName("id")>]
    Id: int

    [<JsonPropertyName("name")>]
    Name: string

    [<JsonPropertyName("email")>]
    Email: string option
}

// Automatic with System.Text.Json or Thoth.Json
let parseUser (json: string) : Result<User, string> =
    try
        JsonSerializer.Deserialize<User>(json) |> Ok
    with ex ->
        Error ex.Message

// Or with Thoth.Json for Elm-style decoders
open Thoth.Json.Net

let userDecoder : Decoder<User> =
    Decode.object (fun get -> {
        Id = get.Required.Field "id" Decode.int
        Name = get.Required.Field "name" Decode.string
        Email = get.Optional.Field "email" Decode.string
    })

Why this translation:

• F# offers two approaches: attribute-based (simpler) or decoder-based (like Elm)
• Thoth.Json provides Elm-style decoders for F#
• System.Text.Json uses attributes and reflection
• Maybe fields → option with proper serialization handling
• F# has more serialization libraries available (.NET ecosystem)

Error Handling

Elm's Error Model

Elm guarantees no runtime exceptions through its type system:

• All errors are encoded in types (Maybe, Result)
• Impossible states are made impossible via union types
• Compiler enforces exhaustive pattern matching

F#'s Error Model

F# has multiple error handling approaches:

1. Option/Result types (recommended for Elm migrations)
2. Exceptions (for interop with .NET libraries)
3. Computation expressions (for error workflows)

Migration Strategy

Preserve Elm's error safety:

// Use Result for expected errors
type ValidationError =
    | EmptyName
    | InvalidEmail
    | AgeTooYoung

let validateUser name email age : Result<User, ValidationError> =
    if String.IsNullOrWhiteSpace(name) then
        Error EmptyName
    elif not (email.Contains("@")) then
        Error InvalidEmail
    elif age < 18 then
        Error AgeTooYoung
    else
        Ok { Name = name; Email = email; Age = age }

// Chain validations
let createUser name email age =
    result {
        let! validatedUser = validateUser name email age
        let! savedUser = saveToDatabase validatedUser
        return savedUser
    }

Handle .NET exceptions when necessary:

// Wrap .NET APIs that throw exceptions
let safeParse (str: string) : Result<int, string> =
    try
        int str |> Ok
    with
    | :? System.FormatException -> Error "Invalid format"
    | :? System.OverflowException -> Error "Number too large"
    | ex -> Error ex.Message

Concurrency Patterns

Elm: Cmd and Tasks

Elm uses Cmd for side effects and Task for asynchronous operations:

type Msg
    = GotUsers (Result Http.Error (List User))

fetchUsers : Cmd Msg
fetchUsers =
    Http.get
        { url = "https://api.example.com/users"
        , expect = Http.expectJson GotUsers usersDecoder
        }

F#: Async and Elmish Commands

F# uses Async<'T> for asynchronous operations and Elmish Cmd<'Msg>:

type Msg =
    | GotUsers of Result<User list, string>

let fetchUsers : Cmd<Msg> =
    Cmd.OfAsync.perform
        (fun () -> async {
            let! response = Http.get "https://api.example.com/users"
            return! parseUsers response
        })
        ()
        (fun users -> GotUsers (Ok users))
        (fun ex -> GotUsers (Error ex.Message))

// Or with computation expression
let fetchUsersAsync () : Async<Result<User list, string>> =
    async {
        try
            let! response = Http.AsyncGet("https://api.example.com/users")
            let! users = parseUsersAsync response
            return Ok users
        with ex ->
            return Error ex.Message
    }

Why this translation:

• Cmd msg → Cmd<'msg> (Elmish)
• Task → Async<'T> (F# async workflow)
• Elmish provides helpers like Cmd.OfAsync.perform
• F# async is more powerful but requires explicit error handling

Common Pitfalls

1. Type Parameter Order in Result

   • Elm: Result error value (error first)
   • F#: Result<'value, 'error> (value first)
   • Always double-check when converting Result types!

2. Constructor Names

   • Elm: Just, Nothing, Err
   • F#: Some, None, Error
   • Remember to rename when converting

3. Record Update Syntax

   • Elm: { model | field = value }
   • F#: { model with Field = value }
   • Different keywords: | vs with

4. Union Case Syntax

   • Elm: Type arg
   • F#: Type of arg
   • F# requires of keyword

5. List vs Array Performance

   • Both Elm and F# list are immutable linked lists
   • F# also has mutable array for performance-critical code
   • Prefer list for Elm semantics, consider array for hot paths

6. Module Qualification

   • Elm: List.map, String.toInt
   • F#: List.map, System.Int32.Parse
   • F# may require fully qualified names for .NET types

7. Null Safety

   • Elm: No null, ever
   • F#: null exists for .NET interop
   • Use Option.ofObj to convert nullable .NET values to option

8. Capitalization Conventions

   • Elm: camelCase for everything except types
   • F#: PascalCase for types and record fields, camelCase for values
   • Must rename fields when converting records

Tooling

| Tool | Purpose | Notes | |------|---------|-------| | Fable | F# → JavaScript compiler | Compile F# to JS like Elm compiles to JS | | Elmish | TEA implementation for F# | Official F# implementation of Elm Architecture | | Elmish.React | React bindings for Elmish | Render views using React | | Thoth.Json | JSON decoding like Elm | Elm-style decoders for F# | | Feliz | Modern F# React DSL | Alternative to Elmish.React | | Ionide | F# IDE support | VS Code extension for F# | | Fantomas | F# code formatter | Like elm-format | | FsCheck | Property-based testing | Like Elm's fuzz testing |

Examples

Example 1: Simple - Type and Function

Before (Elm):

type alias Point =
    { x : Float
    , y : Float
    }

distance : Point -> Point -> Float
distance p1 p2 =
    let
        dx = p1.x - p2.x
        dy = p1.y - p2.y
    in
    sqrt (dx * dx + dy * dy)

After (F#):

type Point = {
    X: float
    Y: float
}

let distance (p1: Point) (p2: Point) : float =
    let dx = p1.X - p2.X
    let dy = p1.Y - p2.Y
    sqrt (dx * dx + dy * dy)

Example 2: Medium - Union Types and Pattern Matching

Before (Elm):

type Tree a
    = Empty
    | Node a (Tree a) (Tree a)

depth : Tree a -> Int
depth tree =
    case tree of
        Empty ->
            0

        Node _ left right ->
            1 + max (depth left) (depth right)

mapTree : (a -> b) -> Tree a -> Tree b
mapTree f tree =
    case tree of
        Empty ->
            Empty

        Node value left right ->
            Node (f value) (mapTree f left) (mapTree f right)

After (F#):

type Tree<'a> =
    | Empty
    | Node of 'a * Tree<'a> * Tree<'a>

let rec depth (tree: Tree<'a>) : int =
    match tree with
    | Empty ->
        0

    | Node (_, left, right) ->
        1 + max (depth left) (depth right)

let rec mapTree (f: 'a -> 'b) (tree: Tree<'a>) : Tree<'b> =
    match tree with
    | Empty ->
        Empty

    | Node (value, left, right) ->
        Node (f value, mapTree f left, mapTree f right)

Example 3: Complex - The Elm Architecture with HTTP

Before (Elm):

module Main exposing (main)

import Browser
import Html exposing (..)
import Html.Events exposing (onClick)
import Http
import Json.Decode as Decode

type alias Model =
    { users : List User
    , loading : Bool
    , error : Maybe String
    }

type alias User =
    { id : Int
    , name : String
    }

type Msg
    = LoadUsers
    | GotUsers (Result Http.Error (List User))

init : () -> ( Model, Cmd Msg )
init _ =
    ( { users = [], loading = False, error = Nothing }
    , Cmd.none
    )

update : Msg -> Model -> ( Model, Cmd Msg )
update msg model =
    case msg of
        LoadUsers ->
            ( { model | loading = True, error = Nothing }
            , fetchUsers
            )

        GotUsers (Ok users) ->
            ( { model | users = users, loading = False }
            , Cmd.none
            )

        GotUsers (Err error) ->
            ( { model | loading = False, error = Just (errorToString error) }
            , Cmd.none
            )

view : Model -> Html Msg
view model =
    div []
        [ button [ onClick LoadUsers ] [ text "Load Users" ]
        , if model.loading then
            text "Loading..."
          else
            div []
                [ viewError model.error
                , viewUsers model.users
                ]
        ]

viewError : Maybe String -> Html Msg
viewError error =
    case error of
        Just err ->
            div [] [ text ("Error: " ++ err) ]

        Nothing ->
            text ""

viewUsers : List User -> Html Msg
viewUsers users =
    ul [] (List.map viewUser users)

viewUser : User -> Html Msg
viewUser user =
    li [] [ text user.name ]

fetchUsers : Cmd Msg
fetchUsers =
    Http.get
        { url = "https://api.example.com/users"
        , expect = Http.expectJson GotUsers usersDecoder
        }

usersDecoder : Decode.Decoder (List User)
usersDecoder =
    Decode.list userDecoder

userDecoder : Decode.Decoder User
userDecoder =
    Decode.map2 User
        (Decode.field "id" Decode.int)
        (Decode.field "name" Decode.string)

errorToString : Http.Error -> String
errorToString error =
    case error of
        Http.BadUrl url ->
            "Bad URL: " ++ url

        Http.Timeout ->
            "Timeout"

        Http.NetworkError ->
            "Network error"

        Http.BadStatus code ->
            "Bad status: " ++ String.fromInt code

        Http.BadBody message ->
            "Bad body: " ++ message

main =
    Browser.element
        { init = init
        , update = update
        , view = view
        , subscriptions = \_ -> Sub.none
        }

After (F#):

module Main

open Elmish
open Elmish.React
open Fable.React
open Fable.React.Props
open Thoth.Json.Decode
open Fable.SimpleHttp

type Model = {
    Users: User list
    Loading: bool
    Error: string option
}

type User = {
    Id: int
    Name: string
}

type Msg =
    | LoadUsers
    | GotUsers of Result<User list, string>

let init () : Model * Cmd<Msg> =
    { Users = []; Loading = false; Error = None }, Cmd.none

let update (msg: Msg) (model: Model) : Model * Cmd<Msg> =
    match msg with
    | LoadUsers ->
        { model with Loading = true; Error = None }, fetchUsers ()

    | GotUsers (Ok users) ->
        { model with Users = users; Loading = false }, Cmd.none

    | GotUsers (Error error) ->
        { model with Loading = false; Error = Some error }, Cmd.none

let view (model: Model) (dispatch: Msg -> unit) =
    div [] [
        button [ OnClick (fun _ -> dispatch LoadUsers) ] [ str "Load Users" ]
        if model.Loading then
            str "Loading..."
        else
            div [] [
                viewError model.Error
                viewUsers model.Users
            ]
    ]

let viewError (error: string option) =
    match error with
    | Some err ->
        div [] [ str ("Error: " + err) ]
    | None ->
        str ""

let viewUsers (users: User list) =
    ul [] (users |> List.map viewUser)

let viewUser (user: User) =
    li [] [ str user.Name ]

let fetchUsers () : Cmd<Msg> =
    Cmd.OfAsync.perform
        (fun () -> async {
            let! response = Http.get "https://api.example.com/users"
            match response.statusCode with
            | 200 ->
                match Decode.fromString usersDecoder response.responseText with
                | Ok users -> return Ok users
                | Error err -> return Error err
            | code ->
                return Error $"Bad status: {code}"
        })
        ()
        GotUsers
        (fun ex -> GotUsers (Error ex.Message))

let userDecoder : Decoder<User> =
    Decode.object (fun get -> {
        Id = get.Required.Field "id" Decode.int
        Name = get.Required.Field "name" Decode.string
    })

let usersDecoder : Decoder<User list> =
    Decode.list userDecoder

Program.mkProgram init update view
|> Program.withReactSynchronous "root"
|> Program.run

See Also

For more examples and patterns, see:

• meta-convert-dev - Foundational conversion patterns with cross-language examples
• lang-elm-dev - Elm development patterns and The Elm Architecture
• lang-fsharp-dev - F# development patterns and functional programming
• patterns-concurrency-dev - Async patterns across languages
• patterns-serialization-dev - JSON handling across languages