Convert Scala to Elm

Convert Scala to Elm

Convert Scala code to idiomatic Elm. This skill extends meta-convert-dev with Scala-to-Elm specific type mappings, idiom translations, and tooling for translating from backend/​full-stack functional programming to frontend functional programming with simpler, safer types.

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: Scala's sealed traits and case classes → Elm's union types
• Idiom translations: Functional Scala patterns → The Elm Architecture (TEA)
• Error handling: Option/​Either with rich combinators → Maybe/​Result with simpler API
• Async patterns: Future/​IO/​Task with effect systems → Cmd/​Sub with TEA
• Type system: Advanced types (higher-kinded, type classes, implicits) → simple types with guarantees

This Skill Does NOT Cover

• General conversion methodology - see meta-convert-dev
• Scala language fundamentals - see lang-scala-dev
• Elm language fundamentals - see lang-elm-dev
• Reverse conversion (Elm → Scala) - see convert-elm-scala
• ScalaJS specific patterns - see lang-scala-js-dev for frontend-to-frontend conversions

---

Quick Reference

---

When Converting Code

1. Analyze source thoroughly before writing target - understand Scala's abstractions and data flow
2. Map types first - create type equivalence table, simplifying advanced features
3. Preserve semantics over syntax similarity - embrace Elm's simpler model
4. Adopt target idioms - don't write "Scala code in Elm syntax"
5. Handle edge cases - simplify Option/​Either to Maybe/​Result, manage effect boundaries
6. Test equivalence - same inputs → same outputs
7. Simplify type system usage - remove higher-kinded types, implicits, type classes

---

Type System Mapping

Primitive Types

Collection Types

Composite Types

---

Idiom Translation

Pattern: Sealed Traits to Union Types

Scala uses sealed traits with case classes/​objects for discriminated unions. Elm uses simpler union types.

Scala:

// Sealed trait ensures exhaustive pattern matching
sealed trait Msg
case object Increment extends Msg
case object Decrement extends Msg
case class SetCount(value: Int) extends Msg

case class Model(count: Int)

def update(model: Model, msg: Msg): Model = msg match {
  case Increment => model.copy(count = model.count + 1)
  case Decrement => model.copy(count = model.count - 1)
  case SetCount(newCount) => model.copy(count = newCount)
}

Elm:

-- Union type with simple syntax
type Msg
    = Increment
    | Decrement
    | SetCount Int

type alias Model =
    { count : Int }

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

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

        SetCount newCount ->
            { model | count = newCount }

Why this translation:

• Elm's union types are simpler and don't require separate declarations for each variant
• Elm's record update syntax { model | count = ... } is more concise than Scala's copy
• Both provide compile-time exhaustiveness checking
• Elm enforces purity at the type system level (no side effects in update)

---

Pattern: Option to Maybe

Scala's Option type with rich combinators translates to Elm's Maybe with simpler API.

Scala:

case class User(name: String, age: Int)

def findUser(id: Int): Option[User] = {
  if (id == 1) Some(User("Alice", 30))
  else None
}

def displayName(maybeUser: Option[User]): String = maybeUser match {
  case Some(user) => user.name
  case None => "Anonymous"
}

// Using Option combinators
val name: String =
  findUser(1)
    .map(_.name)
    .getOrElse("Anonymous")

// Or more idiomatically with fold
val name2: String =
  findUser(1).fold("Anonymous")(_.name)

Elm:

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

findUser : Int -> Maybe User
findUser id =
    if id == 1 then
        Just { name = "Alice", age = 30 }
    else
        Nothing

displayName : Maybe User -> String
displayName maybeUser =
    case maybeUser of
        Just user ->
            user.name

        Nothing ->
            "Anonymous"

-- Using Maybe combinators
name : String
name =
    findUser 1
        |> Maybe.map .name
        |> Maybe.withDefault "Anonymous"

Why this translation:

• Maybe has the same semantics as Option (Just/​Nothing vs Some/​None)
• Elm uses Maybe.withDefault instead of getOrElse
• Elm doesn't have fold on Maybe, use map + withDefault instead
• Elm's pipe operator |> is similar to method chaining
• Property access .name is more concise than Scala's _.name

---

Pattern: Either to Result

Scala's Either type (right-biased) maps to Elm's Result, which is designed for error handling.

Scala:

def parseAge(str: String): Either[String, Int] = {
  try {
    val age = str.toInt
    if (age >= 0) Right(age)
    else Left("Age must be non-negative")
  } catch {
    case _: NumberFormatException => Left("Not a valid number")
  }
}

// Chain Eithers with flatMap
def validateAge(str: String): Either[String, Int] = {
  parseAge(str).flatMap { age =>
    if (age < 120) Right(age)
    else Left("Age must be less than 120")
  }
}

// Or using for-comprehension (idiomatic)
def validateAge2(str: String): Either[String, Int] = for {
  age <- parseAge(str)
  validAge <- if (age < 120) Right(age)
              else Left("Age must be less than 120")
} yield validAge

Elm:

parseAge : String -> Result String 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 number"

-- Chain Results with andThen
validateAge : String -> Result String Int
validateAge str =
    parseAge str
        |> Result.andThen
            (\age ->
                if age < 120 then
                    Ok age
                else
                    Err "Age must be less than 120"
            )

Why this translation:

• Result has the same semantics as Either (Ok/​Err vs Right/​Left)
• Elm uses Result.andThen instead of flatMap for chaining
• Elm doesn't have for-comprehensions; use pipe chains instead
• Elm doesn't have exceptions, so parsing returns Maybe, not throws
• Elm Result is explicitly for error handling, not general sum types

---

Pattern: Functional Effects to The Elm Architecture

Scala's functional effect systems (cats-effect, ZIO) translate to The Elm Architecture's Cmd/​Sub.

Scala (with cats-effect):

import cats.effect.IO

// MODEL
case class Model(count: Int)

def init: Model = Model(0)

// UPDATE
sealed trait Msg
case object Increment extends Msg
case object Decrement extends Msg
case object LoadData extends Msg
case class DataLoaded(data: String) extends Msg

def update(model: Model, msg: Msg): (Model, IO[Unit]) = msg match {
  case Increment => (model.copy(count = model.count + 1), IO.unit)
  case Decrement => (model.copy(count = model.count - 1), IO.unit)
  case LoadData => (model, loadDataEffect)
  case DataLoaded(data) => (model, IO.unit) // handle loaded data
}

def loadDataEffect: IO[Unit] =
  IO(println("Loading data..."))

// VIEW
def view(model: Model): String =
  s"Count: ${model.count}"

Elm:

import Http

-- MODEL
type alias Model =
    { count : Int }

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

-- UPDATE
type Msg
    = Increment
    | Decrement
    | LoadData
    | DataLoaded (Result Http.Error String)

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 )

        LoadData ->
            ( model, loadDataCmd )

        DataLoaded result ->
            case result of
                Ok data ->
                    ( model, Cmd.none )  -- handle loaded data

                Err _ ->
                    ( model, Cmd.none )  -- handle error

loadDataCmd : Cmd Msg
loadDataCmd =
    Http.get
        { url = "https://api.example.com/​data"
        , expect = Http.expectString DataLoaded
        }

-- VIEW
view : Model -> Html Msg
view model =
    div []
        [ text ("Count: " ++ String.fromInt model.count) ]

Why this translation:

• TEA's ( Model, Cmd Msg ) return type is similar to Scala's (Model, IO[Unit])
• Cmd.none is equivalent to IO.unit (no effect)
• Elm enforces the TEA pattern at the type system level
• Effects in Elm are always async and message-based (no synchronous IO)
• Elm runtime manages all effects; you describe what to do, runtime executes

---

Paradigm Translation

Mental Model Shift: Backend/​Full-Stack FP → Frontend FP

Type System Mental Model

---

Error Handling

Scala Error Model → Elm Error Model

Scala approach:

• Multiple error types: Option, Either, Try, exceptions
• Rich combinators: fold, flatMap, recover, recoverWith
• For-comprehensions for sequencing
• Type classes for generic error handling

Elm approach:

• Two error types: Maybe (no error info), Result (with error info)
• Simpler combinators: map, andThen, withDefault
• Pipe chains for sequencing
• Explicit pattern matching for error handling

Translation strategy:

Example:

// Scala
def getUser(id: Int): Option[User] = ???
def getOrders(user: User): Either[String, List[Order]] = ???
def processOrders(orders: List[Order]): Try[Summary] = ???

val result: Either[String, Summary] = for {
  user <- getUser(1).toRight("User not found")
  orders <- getOrders(user)
  summary <- processOrders(orders).toEither.left.map(_.getMessage)
} yield summary

-- Elm
getUser : Int -> Maybe User
getOrders : User -> Result String (List Order)
processOrders : List Order -> Result String Summary

result : Result String Summary
result =
    getUser 1
        |> Result.fromMaybe "User not found"
        |> Result.andThen getOrders
        |> Result.andThen processOrders

---

Concurrency Patterns

Scala Async → Elm Async

Scala concurrency models:

• Future (eager, callback-based)
• IO / Task (lazy, referentially transparent)
• Akka actors (message-passing)
• Parallel execution with parMapN, parTraverse

Elm concurrency model:

• Cmd (command to runtime, always async)
• Sub (subscription to events)
• No parallelism control (runtime manages)
• Everything goes through TEA update loop

Translation strategy:

Example:

// Scala (cats-effect)
def fetchUser(id: Int): IO[User] = ???
def fetchOrders(userId: Int): IO[List[Order]] = ???

val program: IO[String] = for {
  user <- fetchUser(1)
  orders <- fetchOrders(user.id)
} yield s"${user.name} has ${orders.length} orders"

program.unsafeRunSync()

-- Elm
type Msg
    = UserFetched (Result Http.Error User)
    | OrdersFetched (Result Http.Error (List Order))

fetchUser : Int -> Cmd Msg
fetchUser id =
    Http.get
        { url = "/​api/​users/" ++ String.fromInt id
        , expect = Http.expectJson UserFetched userDecoder
        }

fetchOrders : Int -> Cmd Msg
fetchOrders userId =
    Http.get
        { url = "/​api/​orders?userId=" ++ String.fromInt userId
        , expect = Http.expectJson OrdersFetched ordersDecoder
        }

update : Msg -> Model -> ( Model, Cmd Msg )
update msg model =
    case msg of
        UserFetched (Ok user) ->
            ( { model | user = Just user }
            , fetchOrders user.id
            )

        UserFetched (Err _) ->
            ( model, Cmd.none )

        OrdersFetched (Ok orders) ->
            ( { model | orders = orders }, Cmd.none )

        OrdersFetched (Err _) ->
            ( model, Cmd.none )

Why this translation:

• Elm doesn't allow synchronous effects; everything is async via Cmd
• Sequential effects become chained messages in the update loop
• Elm runtime manages all async execution; no manual thread management
• Type safety is preserved through Result types in message variants

---

Advanced Feature Translation

Type Classes to Module Functions

Scala's type classes (via implicits/​given) don't exist in Elm. Use explicit module functions.

Scala:

// Type class
trait Show[A] {
  def show(a: A): String
}

object Show {
  implicit val intShow: Show[Int] = (a: Int) => a.toString
  implicit val stringShow: Show[String] = (a: String) => s""""$a""""

  def show[A](a: A)(implicit ev: Show[A]): String = ev.show(a)
}

// Usage
Show.show(42)        // "42"
Show.show("hello")   // "\"hello\""

Elm:

-- Explicit functions per type
showInt : Int -> String
showInt n =
    String.fromInt n

showString : String -> String
showString s =
    "\"" ++ s ++ "\""

-- No polymorphic show function; use type-specific functions
result1 = showInt 42           -- "42"
result2 = showString "hello"   -- "\"hello\""

Why this translation:

• Elm doesn't have type classes or ad-hoc polymorphism
• Use concrete functions for each type
• This is more explicit but less generic
• Consider using custom types to unify different representations

---

Higher-Kinded Types - Not Supported

Scala's higher-kinded types (type constructors as parameters) don't exist in Elm.

Scala:

trait Functor[F[_]] {
  def map[A, B](fa: F[A])(f: A => B): F[B]
}

def mapTwice[F[_]: Functor, A, B, C](fa: F[A])(f: A => B)(g: B => C): F[C] = {
  val fb = Functor[F].map(fa)(f)
  Functor[F].map(fb)(g)
}

// Works with List, Option, Either, etc.
mapTwice(List(1, 2, 3))(_ + 1)(_ * 2)
mapTwice(Some(5))(_ + 1)(_ * 2)

Elm - Use concrete types:

-- Separate implementations for each type
mapTwiceList : (a -> b) -> (b -> c) -> List a -> List c
mapTwiceList f g list =
    list
        |> List.map f
        |> List.map g

mapTwiceMaybe : (a -> b) -> (b -> c) -> Maybe a -> Maybe c
mapTwiceMaybe f g maybe =
    maybe
        |> Maybe.map f
        |> Maybe.map g

-- Usage
mapTwiceList (\x -> x + 1) (\x -> x * 2) [1, 2, 3]
mapTwiceMaybe (\x -> x + 1) (\x -> x * 2) (Just 5)

Why this translation:

• Elm prioritizes simplicity over abstraction
• Code is more explicit and easier to understand
• No accidental complexity from type-level programming
• Duplicate code is acceptable if it improves clarity

---

Common Pitfalls

1. Trying to use type classes: Elm has no type classes. Use explicit function passing or modules for polymorphism.

2. Expecting exceptions: Elm has no runtime exceptions. All errors must be represented as Maybe or Result.

3. Using mutable state: Elm is purely functional with immutable data. Use record updates and model state instead.

4. Synchronous side effects: All effects in Elm are async via Cmd/​Sub. You can't perform IO directly in functions.

5. Advanced type features: Elm doesn't have higher-kinded types, variance, existential types, or path-dependent types. Keep it simple.

6. For-comprehensions: Elm doesn't have for-comprehensions. Use pipe chains with andThen for monadic sequencing.

7. Lazy evaluation: Elm is strict (eager) by default. Use lazy values explicitly with Lazy.lazy.

8. Tuple size limits: Elm tuples are limited to 3 elements. Use records for larger structures.

9. Dict/​Set key constraints: In Elm, Dict keys and Set values must be comparable (Int, Float, String, etc.). Can't use custom types as keys without workarounds.

10. Missing standard library functions: Elm's standard library is intentionally small. You may need to implement common utilities yourself or use packages.

---

Tooling

Migration strategy:

1. Extract pure business logic from Scala
2. Identify data types and ADTs
3. Map types to Elm equivalents
4. Rewrite logic using Elm idioms
5. Structure as TEA (Model-Update-View)
6. Use elm-test to verify behavior equivalence

---

Examples

Examples should progress in complexity:

Example 1: Simple - ADT Translation

Before (Scala):

sealed trait Status
case object Pending extends Status
case object Active extends Status
case object Completed extends Status

def statusToString(status: Status): String = status match {
  case Pending => "Pending"
  case Active => "Active"
  case Completed => "Completed"
}

val current: Status = Active
println(statusToString(current))  // "Active"

After (Elm):

type Status
    = Pending
    | Active
    | Completed

statusToString : Status -> String
statusToString status =
    case status of
        Pending ->
            "Pending"

        Active ->
            "Active"

        Completed ->
            "Completed"

current : Status
current =
    Active

-- Usage: statusToString current  => "Active"

---

Example 2: Medium - Option/​Maybe and Error Handling

Before (Scala):

case class User(name: String, age: Int)

def findUser(id: Int): Option[User] =
  if (id == 1) Some(User("Alice", 30))
  else None

def validateAge(age: Int): Either[String, Int] =
  if (age >= 18) Right(age)
  else Left("Must be 18 or older")

def getUserAge(id: Int): Either[String, Int] =
  findUser(id)
    .toRight("User not found")
    .flatMap(user => validateAge(user.age))

getUserAge(1)  // Right(30)
getUserAge(2)  // Left("User not found")

After (Elm):

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

findUser : Int -> Maybe User
findUser id =
    if id == 1 then
        Just { name = "Alice", age = 30 }
    else
        Nothing

validateAge : Int -> Result String Int
validateAge age =
    if age >= 18 then
        Ok age
    else
        Err "Must be 18 or older"

getUserAge : Int -> Result String Int
getUserAge id =
    findUser id
        |> Result.fromMaybe "User not found"
        |> Result.andThen (\user -> validateAge user.age)

-- getUserAge 1  => Ok 30
-- getUserAge 2  => Err "User not found"

---

Example 3: Complex - TEA Application with HTTP

Before (Scala with cats-effect):

import cats.effect.IO
import io.circe.generic.auto._
import org.http4s._
import org.http4s.circe._
import org.http4s.client._

case class Post(id: Int, title: String, body: String)

case class Model(
  posts: List[Post],
  loading: Boolean,
  error: Option[String]
)

sealed trait Msg
case object FetchPosts extends Msg
case class PostsFetched(posts: List[Post]) extends Msg
case class FetchFailed(error: String) extends Msg

def update(model: Model, msg: Msg): (Model, IO[Msg]) = msg match {
  case FetchPosts =>
    (model.copy(loading = true), fetchPosts)

  case PostsFetched(posts) =>
    (model.copy(posts = posts, loading = false, error = None), IO.pure(FetchPosts))

  case FetchFailed(error) =>
    (model.copy(loading = false, error = Some(error)), IO.pure(FetchPosts))
}

def fetchPosts(implicit client: Client[IO]): IO[Msg] = {
  val uri = Uri.unsafeFromString("https://jsonplaceholder.typicode.com/​posts")
  client.expect[List[Post]](uri)
    .map(PostsFetched)
    .handleErrorWith(err => IO.pure(FetchFailed(err.getMessage)))
}

def view(model: Model): String = {
  if (model.loading) "Loading..."
  else if (model.error.isDefined) s"Error: ${model.error.get}"
  else s"Posts: ${model.posts.map(_.title).mkString(", ")}"
}

After (Elm):

import Browser
import Html exposing (Html, div, text, button, ul, li)
import Html.Events exposing (onClick)
import Http
import Json.Decode as Decode exposing (Decoder, field, int, string)

-- MODEL
type alias Post =
    { id : Int
    , title : String
    , body : String
    }

type alias Model =
    { posts : List Post
    , loading : Bool
    , error : Maybe String
    }

init : ( Model, Cmd Msg )
init =
    ( { posts = []
      , loading = False
      , error = Nothing
      }
    , Cmd.none
    )

-- UPDATE
type Msg
    = FetchPosts
    | PostsFetched (Result Http.Error (List Post))

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

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

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

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

        Http.Timeout ->
            "Request timeout"

        Http.NetworkError ->
            "Network error"

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

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

-- HTTP
fetchPosts : Cmd Msg
fetchPosts =
    Http.get
        { url = "https://jsonplaceholder.typicode.com/​posts"
        , expect = Http.expectJson PostsFetched postsDecoder
        }

postsDecoder : Decoder (List Post)
postsDecoder =
    Decode.list postDecoder

postDecoder : Decoder Post
postDecoder =
    Decode.map3 Post
        (field "id" int)
        (field "title" string)
        (field "body" string)

-- VIEW
view : Model -> Html Msg
view model =
    div []
        [ button [ onClick FetchPosts ] [ text "Fetch Posts" ]
        , viewContent model
        ]

viewContent : Model -> Html Msg
viewContent model =
    if model.loading then
        text "Loading..."

    else
        case model.error of
            Just error ->
                text ("Error: " ++ error)

            Nothing ->
                ul []
                    (List.map viewPost model.posts)

viewPost : Post -> Html Msg
viewPost post =
    li [] [ text post.title ]

-- MAIN
main : Program () Model Msg
main =
    Browser.element
        { init = \_ -> init
        , update = update
        , view = view
        , subscriptions = \_ -> Sub.none
        }

Why this translation:

• Scala's IO effects become Elm's Cmd in TEA
• HTTP client becomes Http.get with JSON decoder
• Error handling via Result instead of IO error handling
• View is pure HTML (no side effects)
• Elm runtime manages all effects and rendering
• Type safety maintained throughout with no runtime exceptions

---

See Also

For more examples and patterns, see:

• meta-convert-dev - Foundational patterns with cross-language examples
• convert-elm-scala - Reverse conversion (Elm → Scala)
• lang-scala-dev - Scala development patterns
• lang-elm-dev - Elm development patterns

Cross-cutting pattern skills (for areas not fully covered by lang-*-dev):

• patterns-concurrency-dev - Async, channels, threads across languages
• patterns-serialization-dev - JSON, validation, struct tags across languages