Erlang to F# Conversion

Overview

This skill guides the conversion of Erlang code to idiomatic F# while maintaining functional programming principles, concurrent programming patterns, and fault-tolerance capabilities. F# provides strong functional programming support with .NET integration.

Key Language Differences

Type Systems

• Erlang: Dynamic typing with pattern matching
• F#: Static typing with type inference, algebraic data types

Concurrency Models

• Erlang: Actor model with lightweight processes, message passing
• F#: MailboxProcessor (similar to actors), async workflows, Task Parallel Library

Runtime Environment

• Erlang: BEAM VM with hot code swapping, distributed computing
• F#: .NET CLR with async/await, comprehensive standard library

---

Quick Reference

Erlang	F#	Notes
atom()	string or discriminated union	Atoms become strings or DU cases
integer()	int or int64	Fixed-size integers
float()	float	64-bit double
binary()	byte[]	Byte arrays
list()	'a list	Immutable lists
tuple()	'a * 'b	Product types
map()	Map<'k,'v>	Immutable maps
pid()	MailboxProcessor<'T>	Actor-style processing
fun()	'a -> 'b	First-class functions

---

Core Conversion Patterns

1. Module and Function Definitions

Erlang:

-module(calculator).
-export([add/2, multiply/2]).

add(X, Y) -> X + Y.
multiply(X, Y) -> X * Y.

F#:

module Calculator =
    let add x y = x + y
    let multiply x y = x * y

2. Pattern Matching

Erlang:

factorial(0) -> 1;
factorial(N) when N > 0 -> N * factorial(N - 1).

F#:

let rec factorial n =
    match n with
    | 0 -> 1
    | n when n > 0 -> n * factorial (n - 1)
    | _ -> failwith "Negative input not allowed"

3. Records and Types

Erlang:

-record(person, {name, age, email}).

create_person(Name, Age, Email) ->
    #person{name=Name, age=Age, email=Email}.

F#:

type Person = {
    Name: string
    Age: int
    Email: string
}

let createPerson name age email =
    { Name = name; Age = age; Email = email }

4. Lists and Comprehensions

Erlang:

double_list(List) -> [X * 2 || X <- List].
filter_even(List) -> [X || X <- List, X rem 2 =:= 0].

F#:

let doubleList list =
    list |> List.map ((*) 2)

let filterEven list =
    list |> List.filter (fun x -> x % 2 = 0)

5. Actor Model / Process Communication

Erlang:

-module(counter).

start() ->
    spawn(fun() -> loop(0) end).

increment(Pid) ->
    Pid ! {increment, self()},
    receive
        {ok, NewValue} -> NewValue
    end.

loop(Count) ->
    receive
        {increment, From} ->
            NewCount = Count + 1,
            From ! {ok, NewCount},
            loop(NewCount)
    end.

F#:

module Counter =
    type Message =
        | Increment of AsyncReplyChannel<int>

    type CounterAgent() =
        let agent = MailboxProcessor.Start(fun inbox ->
            let rec loop count = async {
                let! msg = inbox.Receive()
                match msg with
                | Increment(reply) ->
                    let newCount = count + 1
                    reply.Reply(newCount)
                    return! loop newCount
            }
            loop 0
        )

        member _.Increment() = agent.PostAndReply(Increment)

    let start() = CounterAgent()

6. gen_server Pattern

Erlang:

-module(my_server).
-behaviour(gen_server).

handle_call(get_count, _From, State = #{count := Count}) ->
    {reply, Count, State};
handle_call({increment, N}, _From, State = #{count := Count}) ->
    NewState = State#{count := Count + N},
    {reply, ok, NewState}.

F#:

module MyServer =
    type Message =
        | GetCount of AsyncReplyChannel<int>
        | Increment of int * AsyncReplyChannel<unit>

    type State = { Count: int }

    type ServerAgent() =
        let agent = MailboxProcessor.Start(fun inbox ->
            let rec loop state = async {
                let! msg = inbox.Receive()
                match msg with
                | GetCount(reply) ->
                    reply.Reply(state.Count)
                    return! loop state
                | Increment(n, reply) ->
                    let newState = { Count = state.Count + n }
                    reply.Reply()
                    return! loop newState
            }
            loop { Count = 0 }
        )

        member _.GetCount() = agent.PostAndReply(GetCount)
        member _.Increment(n) = agent.PostAndReply(fun ch -> Increment(n, ch))

7. Error Handling

Erlang:

safe_divide(_, 0) -> {error, division_by_zero};
safe_divide(X, Y) -> {ok, X / Y}.

F#:

type DivisionError = DivisionByZero

let safeDivide x y =
    match y with
    | 0 -> Error DivisionByZero
    | _ -> Ok (x / y)

// Or using Option
let safeDivide' x y =
    match y with
    | 0 -> None
    | _ -> Some (x / y)

8. Binary Pattern Matching

Erlang:

parse_header(<<Type:8, Length:16, Rest/binary>>) ->
    {Type, Length, Rest}.

F#:

let parseHeader (bytes: byte[]) =
    if bytes.Length < 3 then
        None
    else
        let type' = bytes.[0]
        let length = (uint16 bytes.[1] <<< 8) ||| uint16 bytes.[2]
        let rest = bytes.[3..]
        Some (type', length, rest)

9. ETS Tables to .NET Collections

Erlang:

store(Key, Value) ->
    ets:insert(my_table, {Key, Value}).

lookup(Key) ->
    case ets:lookup(my_table, Key) of
        [{Key, Value}] -> {ok, Value};
        [] -> {error, not_found}
    end.

F#:

module MyTable =
    open System.Collections.Concurrent

    let private table = ConcurrentDictionary<string, obj>()

    let store key value =
        table.[key] <- value

    let lookup key =
        match table.TryGetValue(key) with
        | true, value -> Some value
        | false, _ -> None

---

Common Libraries and Equivalents

Erlang	F# / .NET Equivalent
gen_server	MailboxProcessor, Akka.NET actors
supervisor	Custom supervision, Akka.NET
ETS	System.Collections.Concurrent
httpc	System.Net.Http, FSharp.Data.Http
jsx (JSON)	System.Text.Json, FSharp.Json
cowboy (HTTP)	ASP.NET Core, Giraffe, Suave
lager (logging)	Serilog, NLog

---

Best Practices

1. Embrace Static Typing

• Use F#'s type system to catch errors at compile time
• Define explicit types for domain models
• Use discriminated unions for state machines

2. Preserve Functional Patterns

• Keep functions pure where possible
• Use immutable data structures
• Leverage F# pipeline operators (|>)

3. Adapt Concurrency Models

• Use MailboxProcessor for actor-like behavior
• Consider Akka.NET for complex distributed systems
• Use async workflows for I/O-bound operations

4. Error Handling

• Prefer Option and Result types over exceptions
• Use computation expressions for error propagation
• Reserve exceptions for truly exceptional cases

---

Migration Strategy

Step 1: Analyze Erlang Codebase

• Identify module structure and dependencies
• Map OTP behaviors (gen_server, supervisor)
• Document message-passing patterns

Step 2: Design F# Architecture

• Plan module organization
• Design type hierarchy for records and unions
• Choose concurrency approach

Step 3: Convert Core Logic

• Start with pure functions and data structures
• Convert pattern matching and recursion
• Translate list operations

Step 4: Implement Concurrency

• Replace spawn/receive with MailboxProcessor
• Convert gen_server to agent-based patterns
• Add async workflows for I/O operations

Step 5: Testing and Validation

• Port EUnit tests to FsUnit or xUnit
• Test concurrent behaviors
• Performance testing and optimization

---

See Also

• lang-erlang-dev - Erlang development patterns
• lang-fsharp-dev - F# development patterns
• meta-convert-dev - General conversion methodology
• convert-fsharp-erlang - Reverse conversion