Elm at Prezi

[Ixiaus]

For those interested in compiling down to Javascript with a strongly and statically typed language, there's a brilliant proper subset of Haskell that compiles down to Javascript called "Fay": https://github.com/faylang/fay/wiki

[rtfeldman]

Also Roy, which is statically typed (with type inference) and is designed for JS interop. It's also written in JS and can be installed with just npm install roy http://roy.brianmckenna.org/

[jallmann]

There is also js_of_ocaml, which can be used to compile straight OCaml to Javascript:

http://ocsigen.org/js_of_ocaml/

The site has some cool demos including webgl, canvas, and a OCaml toplevel (REPL).

[zem]

i like the looks of opa too: http://opalang.org/

[revelation]

[..] and every line of JavaScript code costs too much time to maintain.

At this point, go to the next blackboard. Write after me:

You can not debug Elm. There is no debugger. You can not debug Elm. There is no debugger.

[icarus127]

I haven't written any Elm but having written some Haskell I can say it's surprising how often you _don't_ need a debugger in a functional language. I doubt Elm not having a debugger is as difficult as, for instance, Javascript or Coffeescript not having a debugger would be.

[PuercoPop]

Paraphrasing a great programmer: "The best debugger is your brain". I have no issue with people using a debugger, but it is not a must in order to debug, imho.

[triplepoint217]

The bigger problem I encountered trying it out yesterday was that it (at the moment) is fairly happy to generate an empty page and not give any compiler errors.

I assume that will change, but it made it pretty hard for me to get started with the language.

That said, I hope it changes, because Elm looks really neat!

[halacsy]

is there any theoretical limitation?

[egeozcan]

> The cost of writing JavaScript is just too high

Every time I see such strongly worded statements, I get amazed by how the authors got to analyze the skills of every developer in the world.

[subb]

Usually, to calm myself, I reword it like this :

For us/In my opinion, the cost of writing Javascript is just too high.

See? Now the rage is gone! I think I'm going to sell this trick...

[snprbob86]

This is not a statement about the skills of other developers. This is a statement about the applicability of JavaScript to the sorts of problems people are trying to solve with it. There are a lot of nice things about JavaScript that people didn't quite understand or appreciate in the early days, but that doesn't change the fact that the HTML5 stack is deeply flawed in many ways for the needs of application developers. That shouldn't be surprising or controversial: All of these technologies were developed in a simpler time for a simpler set of problems.... and in a rush.

[gonehome]

Here's a list of languages that compile down to Javascript: https://github.com/jashkenas/coffee-script/wiki/List-of-lang....

I didn't realize how many there were.

[SCdF]

It's fairly crazy, but then again there are ~700 programming languages listed on Wikipedia, so really, there are just a lot of programming languages in general.

In terms of Foo->JS there seem to be two main camps, at least that I've noticed: writing a "better" JS (CoffeeScript / Dart et al), and writing something that converts your language to JS so you don't have to learn JS (ClojureScript, Scala.js etc) at all.

The latter can have ~700 of those before you've even started to duplicate yourself, so there is still a ways to go :P

Arguably there are only half a dozen or so "real" contenders in the former category (in the same way when you decide you want to learn a new programming language you don't pick from all ~700).

[zem]

i started a subreddit for people interested in compile-to-js: http://www.reddit.com/r/altjs/

not gained much traction so far, but it's a useful place to post interesting links as i find them.

[zsol]

I recently attended a coding dojo where we wrote code in elm and tests in coffeescript. I was new to both languages but I was surprised how easier elm felt compared to coffeescript (note: I have some haskell background ;]). Here are the results: https://github.com/dzsodzso63/PreziCodingDojo/tree/master/fr...

[zem]

livescript is a more haskellish descendant of coffeescript: http://livescript.net/

[epidemian]

Great news that Evan can dedicate fully to Elm as his main job. I'm very curious about the language and its departure/simplification over traditional DOM manipulation with JS. I hope the best for the project and would love to see some effort for the compiler targeting asm.js :D

[steveklabnik]

I've really enjoyed playing with FRP, I even wrote a little gem in Ruby for it: https://github.com/steveklabnik/frappuccino/

And built some simple GUIs using Shoes: https://github.com/steveklabnik/frp_shoes

[jcurbo]

Me too, I did a project for grad school using FRP. https://github.com/jcurbo/functionally-reactive-solar-system

[aristidb]

It seems like Haskell users are quite intent on not using Javascript... Elm is written in Haskell. So is Fay. And of course ghcjs.

[weego]

The world seems to be intent on writing mountains of code to avoid writing javascript, and usually ending up with a system that compiles to more javascript than if they just wrote javascript. Saying javascript is broken is a nonsense, they mean they feel they should be above writing javascript due to its quirks and inconsistencies. Whatever they write the end result is still the "broken" language they wanted to avoid. It all seems like a huge waste of resources and smart people's time to me, but then I've been writing javascript for getting on for 15 years so perhaps im past it and don't "get it" anymore.

[IanChiles]

I think that may not be the case - it may be more that Haskell is a fabulous language to write compilers and parsers in, and so lots of compilers and parsers are written in it :)

[halacsy]

because JS is flexible, the VMs are really good recently and the web platform is tempting but haskell coders miss safety, succinct syntax and efficiency.

[efnx]

Yes - I think once you start being productive with Haskell you prefer to work in languages with strong type systems that are purely functional (Haskell and subsets of it). When I work in Haskell it feels like a burden has been lifted from my shoulders - for real! Coming back to JS is simply painful.

I think more programmers would experience this if it weren't for Haskell's insane learning curve. It took me a ~year of casual learning and skimming before I could do anything useful. Very casual, though ;)

[fabokzs]

I'm wondering if we need a different testing approach for these new wave "functional programming languages".

[jessedhillon]

I find function application without parentheses enclosing the parameter list to be highly unreadable. Is there something wrong with me?

[epidemian]

I think i also got confused the first time i saw that syntax (being used to C-like languages hehe). It started making lots of sense once i got familiar with the concept of currying and functions as first-class values.

Suppose you have a function f with type Int -> Int -> Int (you can read it as "it takes two integers and returns an integer" or "it takes one integer and returns a function that takes an integer and returns an integer"), then:

    f

is the function itself, and it is a usable value.

    f 42

is a function of type Int -> Int, and a perfectly valid value too. And

    f 42 33

is just and Int.

It's like with each application you take one arrow from the function type; until you get something that's not a function any more :)

[tikhonj]

It's just new and you're not used to it. There's nothing wrong with finding new and different syntax confusing, as long as you're willing to keep an open mind.

If you use it for a while, you'll get used to it. After a while, I've found it makes the most sense. I think it's better because it reduces the amount of syntactic noise in the code.

It's a simple design decision. Function application is the most important part of the language. So it gets the best syntax: none.

[sixbrx]

A lot of modern mathematics also drops the parens, especially in areas related to modern algebra. It really makes the most sense and is less ambiguous, with parens not being overloaded between grouping and application.

[efnx]

It does take getting used to. Also, seeing $ everywhere on a line and . not being property access. Though with lenses...I won't get into that...

[d0m]

Lol, nothing wrong with you. Just not used to it.

[Dewie]

I've preferred f x y over f(x,y) ever since I first saw it used in Haskell. Maybe it's just a matter of preference.

[egonschiele]

I am so excited! It's nice to see a functional language like Elm getting financial support from a company.

[AndreasFrom]

Elm 0.8 was recently announced with a range of improvements for those interested: http://elm-lang.org/blog/announce/version-0.8.elm

[tingletech]

not to be confused with the elm, the old unix mail client:

http://www.instinct.org/elm/

[obviouslygreen]

I knew this was unlikely to be the subject of the article but clicked through anyway... oh well.

[205guy]

I'm not the only one. Hey, you can still run pine if you want.

[jlatt]

I was really impressed by elm, but I strongly prefer working with a library that provides some of the features without transpiling. If anyone else is interested, I'm looking for some help with my project: https://github.com/jlatt/frp.js

[j_s]

See also Microsoft Research's Reactive Extensions for JavaScript, released under the Apache License 2.0:

https://github.com/Reactive-Extensions/RxJS

[Nekorosu]

Elm is the future and I'm glad to see it now has some good company's support.

[cuttooth]

Elm is the future as much as Java is dead in the software development industry.

[saosebastiao]

Awesome...you have me intrigued. I would worship the ground you walk on if you targeted asm.js.

[boothead]

Slightly off topic but does anyone know of ans FRP library for python?

[nahname]

Inversion of control is a good thing. Why would you want to avoid it?

[laszlok]

I think the idea here is that you aren't /forced/ into inversion of control. That is, if it is clearer to write things side by side, instead of separated into another callback function, you can keep them together.

[icarus127]

I'm genuinely curious, why do you think inversion of control is an unqualified Good Thing™?

[zeckalpha]

It's not avoided, it just has a different syntax. (Monadic rather than callbacks, but since they are the same thing...)

[jluan]

Congratulations, Evan!

[wheatBread]

Ahhhh, I recognize that user name! Thank you!! :D