Hacker News Reader

Comment by _0w8t | Hacker News Reader

PureScript is a fully static strict functional language with a type system that is more advanced than that in Haskell as it allows to model all JS side-effects. It is based on the ideas for the type system in Koka [1]. That in turns allows more easily to integrate PureScript with JavaScript than Elm, but the language itself is anything but a thin layer over JavaScript.

[1] - http://research.microsoft.com/en-us/projects/koka/

View on HN

Why Koka?

This is not snark. Why Koka when ML and Haskell exist? What does it bring to the table that is new? What Haskell calls monads Koka calls effects? But Haskell has lazy evaluation, right -- and Koka is strict (like ML) and tracks effects (like Haskell). Does Koka complete this reductive functional programming matrix?

     / Monads  / No monads†
   -----------------------
   S | Koka    | ML
   L | Haskell | Miranda

† No monads == Nomads?

In a lazy language like Haskell, you often define pure functions whose result is an effectful monadic action. In a strict language, it makes more sense to qualify functions themselves as being pure or effectful. So you end up working with Kleisli arrows, which are still related to monads.

Also, the way Haskell handles composing effects (monad transformers), while precise, sometimes leaves a lot to be desired in terms of usability. That's why other approaches to managing effectful computations are being explored, like algebraic effects.

Koka tags effects using something like row types. This may very well be easier to use. There's no reason Haskell couldn't do this if it had row types, but it doesn't, so it doesn't.

My google-fu is not turning up any definition of `row type'. Are we talking record types here? Probably not because doesn't Haskell have a record type (much maligned though it is).

You can think of structural record types as being formed by applying a special type constructor to a type-level collection of types:

    row Fields = [name: String, age: Int]
    type Person = MkRecord Fields

Which have kinds:

    MkRecord :: Row Type -> Type
    Fields :: Row Type
    Person :: Type

You can construct variant types in a similar way:

    row Cases = [success: SyntaxTree, failure: ParseError]
    type ParseResult = MkVariant Cases

Which have kinds:

    MkVariant :: Row Type -> Type
    Cases :: Row Type
    ParseResult :: Type

Rows may contain things other than proper types, for example:

    row Effects = [foo: State s, bar: Except e]
    type Action = MkFunction Effects

Which have kind:

    -- function types now take three arguments:
    --  - row of effects
    --  - argument type
    --  - return type
    MkFunction :: Row Effect -> Type -> Type -> Type
    Effects :: Row Effect
    Action :: Type -> Type -> Type

I'm going to go with:

   row = sequence
   record = named conjunctive sequence
   variant = named disjunctive sequence

Can you explain what you mean when you say "proper" types? (So I can get what you mean from there on out.) Are all your examples using regular standard Haskell syntax? (I get that you are making up type (and kind) names for illustration purposes). Thanks!

I got a little bit creative with syntax. Haskell doesn't have rows at all, and its kind of proper types is called “*”, rather than “Type”. Other than that, the syntax is pretty similar.

A proper type is one that has kind `Type` (or `*` in Haskell).

Yes, "record type" is probably better way of describing it. I think "row type" might be used when using record types in the context of database tables.

Interesting :)

When I was gathering information about typed languages that compile to JS, I found Elm rather nice, but calling JS libs was rather cumbersome. Somewhere I read that PureScript has a better integration with JS, so I thought it was a "thinner" layer :D

Purescript doesn't have a runtime, unlike Elm, so in that sense it may be a thinner layer. However, the language itself is more advanced than Elm, allowing for more powerful abstractions but also having the side effect (haha) of having the same steep learning curve as Haskell.

Also, Purescript's FFI is really easy to use (unlike Elm's), afaik this was one of its original design goals.

That sounds very interesting, I'll check out the modelling of those side effects, that's something I would love to understand. Thanks.

One interesting application of effect types ala Koka/PureScript is that they can model type-safe private exceptions that can only be caught by code that have full access to implementation. In Haskell exceptions are unsafe and can trivially be used to subvert the type system.

> In Haskell exceptions are unsafe and can trivially be used to subvert the type system.

That's news to me. Could you post a proof?

https://existentialtype.wordpress.com/2012/08/14/haskell-is-...

Firstly, it's been fixed. Secondly, it wasn't to do with exceptions but with Typeable.

Reply on news.ycombinator.com