06:35 <mkfort> hi, anyone here? I have a question

06:35 <mkfort> is there a way to make 2 mutually recursive variant types? This this make sense?

06:36 <mkfort> ah never mind

06:36 <mkfort> I type mutually recursive variant types into google and I find out how to do this, nice

06:41 <mkfort> well I'm still having a little trouble so if anyone is here it would be nice

06:42 <merriam_> i'm here, but i probably can't help much

06:42 <mkfort> oh I just mispelled

06:42 <mkfort> I'm figuring it out

06:42 <merriam_> ah :)

06:42 <mkfort> hehe, those spelling mistakes kill ya

06:43 <mkfort> i'm doing type state = State of string * transitions and transitions = Empty | Transitions of char * state *transitions;;

06:43 <mkfort> ml is so nice

06:44 <mkfort> so now that I've got my data structure thats half the work right

06:44 <mkfort> time for a cigarrette hehe

06:45 <mkfort> this is going to blow away my theory of computation class, they are using C++ but screw that

06:51 <mkfort> I don't see how anyone can *think* in c++

06:52 <merriam_> a us student using ocaml -- quite rare, i think :)

06:52 <mkfort> it like you can either take 3 pages to write some ugly ass hack that doesn't do anything or you can take less than a page to write a solution that *solves the problem*

06:52 <merriam_> skylan looks possibly usan, but then it's just you and him here

06:52 <mkfort> well I don't think my professor has ever heard of it but he will learn hehe

06:53 <mkfort> caml is the best thing the french have ever done, take ML, which is quite possibly one of the best languages , and then extend it, and then write completely awesome tools and compilers, its incredible. I mean ML should be slow and its a feat of man that they have made something so incredible

06:54 <mkfort> that can actually be *faster* than C, it boggles the mind

06:55 <merriam_> not sure about "should be slow", but i agree its speed is impressive

06:55 <mkfort> well of course I'm not serious but people expect functional languages to be slower, if nothing else because its very high level

06:57 <mkfort> caml is much more expressive than Java and thats slow as beans and its made by some of the largest computer companies in the world

06:58 <mkfort> ah, you are Deutsch, I'm studying German now, but its just like my 2nd week in the class. Another structured but expressive language so to say

06:59 <merriam_> actually i'm english, and i don't know much german :)

06:59 <mkfort> Warum, Warum, Warum. Warum ist das Banana krumm? Darum

06:59 <mkfort> ah

06:59 <merriam_> but i understand that :)

06:59 <mkfort> well english is a nice language but its more like perl don't you think haha

07:01 <mkfort> well thats about all I understand

07:01 <merriam_> learners of german should read this -- http://www.crossmyt.com/hc/linghebr/awfgrmlg.html

07:01 <mkfort> nice, my prof was talking about that actually, I'll definitely read it

07:01 <mkfort> thanks

07:02 <merriam_> especially "the Tale of the Fishwife and its Sad Fate" :)

07:03 <mkfort> at any rate it is much more regular than english, but thats not saying anything

07:17 <mkfort> do you know how I can set 2 vals that are mutually recursive in a lazy way so that it doesn't evaluate the recursion when its defined?

07:19 <mkfort> maybe I'm not thinking about this right

07:19 <mkfort> I see

07:22 <mkfort> just have the transition hold a string I think not the whole thing

07:22 <mkfort> val q1 : state =

07:22 <mkfort> State ("q1", Transition ('0', "q1", Transition ('1', "q2", Empty)))

07:22 <mkfort> so forth

07:25 <mkfort> yes I think this should be able to work

08:01 <mrvn_> mkfort: you can use mutable or references.

08:02 <mkfort> I see

08:10 <mkfort> This expression has type state ref but is here used with type state ref

08:10 <mkfort> I do not understand this error

08:11 <mkfort> I made the type state ref, and then when I used the Transition constructor i did ref q1 (q1 is a state)

08:11 <mrvn_> no_maam: prosper kann nach und nach Punkte zu nem Slide hinzufuegen.

08:11 <mrvn_> ups

08:12 <mkfort> but it seems that 'ref q1' is of type state ref

08:12 <mkfort> which is what the error says but why does it error?

08:12 <mrvn_> paste the error

08:14 <mkfort> # let q1 = State ("q1",Transition('0', ref q1, Transition('1', ref q2, Empty))) and q2 = State ("q2", Transition('1', ref q2, Transition('0', ref q1, Empty)));;

08:14 <mkfort> This expression has type state ref but is here used with type state ref

08:14 <mkfort> that is the error

08:14 <mkfort> makes little sense to me

08:14 <mkfort> I want this thing but I want the same thing

08:14 <mkfort> type state = State of string * transition

08:15 <mkfort> and transition = Empty | Transition of char * state ref * transition;;

08:15 <mkfort> that is the type declaration

08:16 <mrvn_> let rec q1 ... ?

08:16 <mrvn_> Otherwise your ref q1 is for the old q1 and not the onw your defining.

08:17 <mkfort> hmm, if I do that, I get a big recursion

08:17 <mkfort> which is what I was trying to avoid with the references

08:17 <mrvn> You want recursion.

08:18 <mkfort> okay, but this isn't infinite?

08:18 <mrvn> no, its circular.

08:18 <mkfort> okay, its just that when I type that it fills up the screen with the recursion, which is maybe what I want

08:19 <mkfort> but doesn't seem quite right to do that

08:19 <mrvn> shouldn't fill up anything.

08:19 <mkfort> it is evaluating the recursion

08:20 <mrvn> Thats just the output.

08:21 <mkfort> okay, that just sort of through me off, I thought it was like an error without being an error

08:21 <mkfort> danke

08:21 <mkfort> threw me off

08:21 <mrvn> The pretty_print follows the reerences and never sees the end.

08:22 <mkfort> ah, Ich verstehe

08:22 <mkfort> I think that my original way is probably okay then

08:22 <mrvn> works without the ref too.

08:23 <mrvn> But then you have to define it all in one big let rec.

08:23 <mkfort> yes, thats what I want, the pretty_print just confused me because I was testing things out in the interactive mode just to make sure my syntax was correct

08:23 <mkfort> and it doesn't look good to scroll the screen for 300 lines but I see that it doesn't matter

08:26 <mrvn> confusing.

08:41 <mkfort> oh wow, I didn't realize that the entire o-reilly book is online, that makes it much easier

08:47 <mrvn> mkfort: url?

14:26 <xtrm> is there an irc ocaml lib ?

14:47 <xtrm> somewhere?

23:30 <urth^> does anyone know anything about the suitability of ocaml for implementing a neural network?