<Smerdyakov>
dark_light, are you familiar with streams in OCaml and trying to lead him to discovering his own problem?
<BetaTEST>
i wonder :)
<dark_light>
Smerdyakov, no
<dark_light>
Smerdyakov, but i like to learn anyway
<Smerdyakov>
dark_light, then I think you are anti-helping. ;)
<dark_light>
Smerdyakov, well:(
Bigbang is now known as Bigb[a]ng
<BetaTEST>
it's interesting that the same code works fine in caml light (except some syntactical modifications)
<Smerdyakov>
BetaTEST, Caml Light is only meant for education... it is designed/maintained with entirely different considerations.
<Smerdyakov>
You're in France, right?
<BetaTEST>
that's why i want to use ocaml :)
<BetaTEST>
i'm in russia
<Smerdyakov>
Oh. Add the 2nd country to the list of places with bozos still using Caml Light, I guess. ;)
<BetaTEST>
probably, but the fact is - it works.. :)
<Smerdyakov>
SML works better.
<BetaTEST>
than ocaml?
<Smerdyakov>
I can certainly understand hesitation to use a slipshod abomination like OCaml in a classroom, but SML doesn't suffer from the same ad-hoc design.
<BetaTEST>
i'm not in a classroom. i'm doing a research and is interested in language that will satisfy my needs: composable types and some hints to write a compiler
<Smerdyakov>
"Composable types"?
<BetaTEST>
with constructors that can be matched
<Smerdyakov>
I still don't understand. Can you give an example
<BetaTEST>
type a = A a | B a | C | D a ...
<Smerdyakov>
So you just mean regular variants?
<BetaTEST>
maybe :)
<Smerdyakov>
OK. Every ML language has those.
<Smerdyakov>
What makes you choose to use streams instead of ml-yacc?
<BetaTEST>
compile won't be center of the program
<BetaTEST>
+r
<Smerdyakov>
I don't understand.
<BetaTEST>
okay, the program will deal with some model
<BetaTEST>
some kind of compositional model
<Smerdyakov>
Why does that mean that it's better to use streams than ocamlyacc?
<BetaTEST>
i didn't seen ocamlyacc yet :)
<BetaTEST>
*haven't*
<Smerdyakov>
All of the parsers I've written for OCaml use ocamlyacc.
<BetaTEST>
doesn't matter -- the "programming language" i will be using is as simple as calculator's, i don't think it's worth using ocamlyacc
<Smerdyakov>
Well, I can't help you with streams. Maybe someone else can.
<BetaTEST>
hmm... does SML have some kind of "type classes" that Haskell has?
<Smerdyakov>
No
<BetaTEST>
i really like that part of haskell
<Smerdyakov>
Why not use Haskell for your project?
<BetaTEST>
monads :)
<Smerdyakov>
I assume that in particular you mean the requirement of writing imperative code with the IO monad.
<BetaTEST>
IO and others
<Smerdyakov>
Which other aspects of monads are _negative_ features of Haskell?
<BetaTEST>
i'm not saying it's negative, that's being investigated :)
<Smerdyakov>
Which other aspects of monads are reasons for not using Haskell for your project?
<BetaTEST>
my project is to make some modelling extension
<BetaTEST>
i'm planning the possibility to integrate it in some other applications
<BetaTEST>
.net, unix, everything, ocaml is better for that than haskell (as far as i know)
<Smerdyakov>
Probably true
<BetaTEST>
it's interesting, when do microsoft plan to adopt Haskell in some kind of H# :)
<BetaTEST>
i don't think f# will survive
<Smerdyakov>
How is that? It's getting increasingly popular.
<BetaTEST>
f#? last time i tried it even examples didn't compile
<Smerdyakov>
I've used it extensively for a complete project inside Microsoft.
<Smerdyakov>
It's used in some shipping MS products today.
<BetaTEST>
okay, i'll try new version
<BetaTEST>
so f# is stable enough to be shipped with production-quality software?
<Smerdyakov>
Yes.
<Smerdyakov>
Has been for years
<Smerdyakov>
Since my experience was inside MS, I can't really say how long the versions I'm talking about have been publicly available.
hikozaemon has joined #ocaml
<BetaTEST>
hmm.. does it support streams?
<BetaTEST>
looks like it does, but [< >] doesn't work
<BetaTEST>
hmm.. those lazy lists are kinda interesting