CommunityNews

CommunityNews

Tail-call optimization in Elm

What is TCO?

Tail-call optimization (TCO) is a very neat trick that the Elm compiler does to make recursive functions a lot more performant and stackoverflow-proof.

Evan Czaplicki describes it very well in this article and I recommend you go read it. He calls it tail-call elimination but it’s a different name for the same thing.

To summarize Evan’s article, a “tail-call optimized” function is a recursive function that gets compiled to using a loop instead of function calls to itself. Let’s take the following code as an example…

Read in full here:

This thread was posted by one of our members via one of our news source trackers.

Most Liked

OvermindDL1

OvermindDL1

That’s not TCO that Elm does, it’s self-call tail recursive optimization, it only can happen in fully known contexts and is not general enough to be full TCO. Full Tail Call Optimization is where a call is in the tail-most position of a function, that means you can then compile essentially a goto to it (or in the Rust proposal a become for example), this means the call can be dynamically made, can be any amount deep through any amount of function stacks, etc… etc… I really don’t like how the Elm community keeps trying to misrepresent things about the Elm language, it’s an annoying recurring pattern…

In fact, let’s try it, I go to: Try Elm!
And I try running:

import Html exposing (text)

entry f i acc = if i <= 0 then acc else f (i - 1) (acc + i)

dispatch1 i acc = entry dispatch2 i acc

dispatch2 i acc = entry dispatch1 i acc

main =
  let i = dispatch1 10000 0 in
  text (String.fromInt i)

And the result is:

Initialization Error

InternalError: too much recursion

Well, let’s try this in Elixir, a language that DOES support TCO, so the same code ported running through the repl:

❯ iex
Erlang/OTP 24 [RELEASE CANDIDATE 3] [erts-12.0] [source] [64-bit] [smp:16:16] [ds:16:16:10] [async-threads:1] [jit]

Interactive Elixir (1.12.0-rc.1) - press Ctrl+C to exit (type h() ENTER for help)
iex(1)> defmodule TestingTCO do
...(1)>   def entry(f, i, acc), do: if(i <= 0, do: acc, else: f.(i-1, acc+i))
...(1)>   
...(1)>   def dispatch1(i, acc), do: entry(&dispatch2/2, i, acc)
...(1)>   
...(1)>   def dispatch2(i, acc), do: entry(&dispatch1/2, i, acc)
...(1)> end
{:module, TestingTCO,
 <<70, 79, 82, 49, 0, 0, 7, 176, 66, 69, 65, 77, 65, 116, 85, 56, 0, 0, 0, 202,
   0, 0, 0, 20, 17, 69, 108, 105, 120, 105, 114, 46, 84, 101, 115, 116, 105,
   110, 103, 84, 67, 79, 8, 95, 95, 105, 110, ...>>, {:dispatch2, 2}}
iex(2)> TestingTCO.dispatch1(10000, 0)
50005000

And since Elixir really does implement TCO and not just a simple recursive loop optimization unlike elm’s communities repeating lies then we can go way way higher!

iex(3)> TestingTCO.dispatch1(1000000, 0)
500000500000

Elm has a lot of issues both as a language and as a community, and passing off a very very common optimization pass performed in almost every language as TCO is just the tip of the iceberg…

AstonJ

AstonJ

Even if it’s not true TCO, performance enhancements are always welcome when it comes to JS land :nerd_face:

rustkas

rustkas

I have already read a lot about the disadvantages of Elm. In your opinion, what is currently better than Elm (in terms of the quality of code generation, what to choose for creating a new application) for the front end part?

Where Next?

Popular Frontend topics Top

New
First poster: AstonJ
Welcome to the 5th edition of the JavaScript Rising Stars, our annual round-up of the JavaScript landscape! The concept is the same as b...
/js
New
First poster: bot
Libsodium has been fully supporting WebAssembly as a target for quite a long time. This includes its built-in benchmark suite, that can r...
New
First poster: dimitarvp
The future of web-based software architectures is already taking form, and this time it’s server-rendered (again). Papa’s got a brand new...
New
First poster: dyowee
The key reason why you (and likely most developers) struggle with CSS, is that you underestimate it. Underestimating CSS leads to a stra...
New
First poster: bot
In this article, I will discuss my journey from being an anti-TypeScript developer to a developer who now couldn’t think of going back to...
New
First poster: claudio
You’re at a restaurant, and there’s an odd item on the menu that you’ve never heard of before, but it piques your interest. It sounds lik...
New
First poster: bot
Humio software engineers Thomas Anagrius and Jeroen Engels sat down to talk about why they got involved with Elm for web-based front-end ...
New
First poster: bot
Since the Humio web client is built in Elm, I’d like to share some of our learnings with Elm over the years. Specifically, working with U...
New
First poster: bot
Recently Tom MacWright has written a few posts on Single Page Applications and their discontents: Second-guessing the modern web If not...
New

Other popular topics Top

PragmaticBookshelf
Machine learning can be intimidating, with its reliance on math and algorithms that most programmers don't encounter in their regular wor...
New
Rainer
My first contact with Erlang was about 2 years ago when I used RabbitMQ, which is written in Erlang, for my job. This made me curious and...
New
AstonJ
Continuing the discussion from Thinking about learning Crystal, let’s discuss - I was wondering which languages don’t GC - maybe we can c...
New
foxtrottwist
A few weeks ago I started using Warp a terminal written in rust. Though in it’s current state of development there are a few caveats (tab...
New
New
DevotionGeo
I have always used antique keyboards like Cherry MX 1800 or Cherry MX 8100 and almost always have modified the switches in some way, like...
New
First poster: bot
zig/http.zig at 7cf2cbb33ef34c1d211135f56d30fe23b6cacd42 · ziglang/zig. General-purpose programming language and toolchain for maintaini...
New
New
PragmaticBookshelf
Use advanced functional programming principles, practical Domain-Driven Design techniques, and production-ready Elixir code to build scal...
New
xiji2646-netizen
Woke up to this today: Claude Code’s complete source code exposed via npm source map. Not a snippet. All 512,000 lines. 1,900 TypeScript ...
New