18:07 <gzar> Whats the core difference between haskell and ocaml?

18:11 <bartholin> gzar: haskell needs monads to get side-effects, in ocaml you can get side-effect in a simple way, like in an imperative language

18:13 <gzar> you can have monads in ocaml as well, but besides that, i hardly could tell whats different

18:13 <gzar> so its the imperative features that differentiate it

18:16 <gzar> thats very cool. It kinda bugs me a bit that i can't really use it anywhere though

18:21 <bartholin> also, Haskell has typeclasses, and ocaml has objects, but those are advanced features

18:23 <gzar> dont know what typeclasses are, but i familiarized myself with ocamls objects a little

19:40 <discocaml> <yawaramin> many people are using ocaml in production https://ocaml.org/industrial-users

20:44 <discocaml> <functionalprogramming> typeclasses vs ML modules is the largest difference i’d say

20:52 <gzar> is the syntax also different? from the looks of it its only vaguely simmilar

20:52 <discocaml> <Kali> ocaml is imperative, object-oriented, and functional, while haskell is pure functional

20:53 <discocaml> <Kali> their methods of semantic organization also differ (ocaml modules, haskell typeclasses)

20:53 <discocaml> <Kali> ocaml's language features are non-optional besides a select few while haskell permits you to enable and disable them at will

20:55 <discocaml> <Kali> haskell is lazy and ocaml is strict (evaluation)

20:55 <gzar> alright, thanks, that clarifies it

20:55 <discocaml> <Kali> their syntax is fairly different but there are many parallels

20:56 <discocaml> <Kali> ocaml ignores whitespace and haskell uses it semantically

20:57 <discocaml> <Kali> ocaml has algebraic effects and haskell does not

20:57 <discocaml> <Kali> those are the major differences

21:08 <gzar> typeclasses seem to be just variant types, or am i wrong on this

21:09 <discocaml> <yawaramin> no they are more like interfaces on a superficial level

21:09 <discocaml> <yawaramin> OCaml functions also have named and optional arguments

21:10 <gzar> yeah but the named and optional arguments really mess with the type system

21:19 <discocaml> <deepspacejohn> I don't think they have any effect on the type system? other than that function types can have named and optional arguments in their signatures in addition to positional arguments.

21:20 <discocaml> <yawaramin> there's a recommended way of using them which works well

21:21 <discocaml> <yawaramin> oh and huge difference–compile speeds. OCaml is similar to Go (fast), Haskell is similar to Rust (slow)

21:21 <discocaml> <yawaramin> personally, fast iteration loop is super important for me

23:05 <gzar> true yeah. what i mean is, if you define an ocaml function with a named parameter, you cannot use that function without explicitly stating the named parameter. you cant omit the name. And it has a different signature, meaning even if 2 functions take and return an int, if one of them uses a named parameter, you cannot use them interchangibly

23:15 <discocaml> <yawaramin> > you cant omit the name

23:15 <discocaml> <yawaramin> that's desirable

23:30 <gzar> maybe, but the type-incompatbility between the functions isnt for me

23:30 <gzar> a parameter with a different name will produce a different function signature

23:37 <discocaml> <yawaramin> sure, people have different preferences, and can even change with time 🤷‍♂️

23:38 <discocaml> <leviroth> I think it would be really annoying and error-prone if the following type-checked: `module M : sig val send_email : require_confirmation:bool -> unit end = struct let send_email ~skip_confirmation = ... end`

23:41 <gzar> fair point

23:47 <discocaml> <contextfreebeer> can't be overstated how fast the compiler is, it's insane

23:48 <discocaml> <contextfreebeer> the compiler is doing a lot less than e.g. ghc but still