2012-04-18T00:04:01 *** thestinger has joined #aichallenge 2012-04-18T00:10:46 *** amstan has quit IRC (Quit: Konversation terminated!) 2012-04-18T00:55:18 *** thestinger has quit IRC (Quit: sleep(28800)) 2012-04-18T01:04:10 *** replore_ has joined #aichallenge 2012-04-18T01:07:40 *** Areks has joined #aichallenge 2012-04-18T01:20:40 *** Accoun has quit IRC (Ping timeout: 244 seconds) 2012-04-18T01:27:58 *** kilae has joined #aichallenge 2012-04-18T02:00:08 *** Fandekasp has quit IRC (Quit: Lost terminal) 2012-04-18T02:02:52 *** Fandekasp has joined #aichallenge 2012-04-18T02:21:37 *** replore_ has quit IRC (Remote host closed the connection) 2012-04-18T02:41:23 *** pairofdice has joined #aichallenge 2012-04-18T02:47:38 *** Palmik has joined #aichallenge 2012-04-18T02:55:24 *** Vrexu has quit IRC (Remote host closed the connection) 2012-04-18T03:04:06 *** Vrexu has joined #aichallenge 2012-04-18T03:06:07 *** kurnevsky has joined #aichallenge 2012-04-18T03:11:18 *** Garf has joined #aichallenge 2012-04-18T04:07:11 *** kurnevsky has quit IRC (Quit: Leaving.) 2012-04-18T04:08:06 *** Fandekasp has quit IRC (Quit: leaving) 2012-04-18T04:40:18 *** mceier has joined #aichallenge 2012-04-18T04:52:23 *** delzex has joined #aichallenge 2012-04-18T04:54:52 *** delzex has quit IRC (Client Quit) 2012-04-18T05:18:42 *** Vrexu has quit IRC (Remote host closed the connection) 2012-04-18T05:33:55 *** mcstar has joined #aichallenge 2012-04-18T06:03:20 mcstar: http://forum.dlang.org/thread/jlvge6$trk$1@digitalmars.com?page=3#post-mailman.1649.1334228692.4860.digitalmars-d:40puremagic.com :D 2012-04-18T06:03:41 *** Areks has quit IRC (Ping timeout: 272 seconds) 2012-04-18T06:04:33 now how do you read from a file in haskell ? 2012-04-18T06:04:58 mleise: hi 2012-04-18T06:05:11 hi 2012-04-18T06:05:50 it is easy to read from a file, what do you mean? 2012-04-18T06:06:03 readFile "file" 2012-04-18T06:09:39 oh ... that's really simple. you don't need to explain category theory first it seems 2012-04-18T06:09:49 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2012-04-18T06:10:04 no i dont 2012-04-18T06:10:39 mleise: ofc, this is a recommended read http://www.haskell.org/haskellwiki/Typeclassopedia 2012-04-18T06:11:03 ah, i think you already knew that 2012-04-18T06:11:39 Yes, and for an outsider that is just overwhelming. 2012-04-18T06:12:18 uhhh 2012-04-18T06:12:28 Even with good explanations it is often difficult to see the benefit of changes. Like what is partial function application? 2012-04-18T06:12:45 "It gets worse when you go: "How does this help 2012-04-18T06:12:49 read from a file" and they give you a long stare and start all over 2012-04-18T06:12:51 again, I just want to know how to read from a goddamn file!" 2012-04-18T06:12:53 ^^^ rofl 2012-04-18T06:13:04 hehe 2012-04-18T06:13:10 :D 2012-04-18T06:13:21 mleise: you kidding right? you must know what is partial function application 2012-04-18T06:13:34 one thing i hate is CL is that you need a lambda for that 2012-04-18T06:13:40 no, I'm not. guess why i never touched FP 2012-04-18T06:14:09 *** foRei has joined #aichallenge 2012-04-18T06:14:13 it's all so mathsy 2012-04-18T06:14:13 mleise: even c++ guys use that all the time 2012-04-18T06:14:27 (i know i did when i wrote c++) 2012-04-18T06:16:13 mleise: do you really argue with, that this file is readable? http://hpaste.org/67201 2012-04-18T06:16:21 *** Areks has joined #aichallenge 2012-04-18T06:17:16 (this code does flying objects and collision detection, so i called it asteroids, though, it is far from that) 2012-04-18T06:17:19 Ship Double Double Double Double Double 2012-04-18T06:17:37 x y orinetation vx vy 2012-04-18T06:17:47 mcstar: what does "data FColor = .... dereived ( ... )" mean? 2012-04-18T06:17:54 derived 2012-04-18T06:18:16 dereiving 2012-04-18T06:18:17 pairofdice: i oculd have used record syntax, but i wanted to keep the data declarations in the same file 2012-04-18T06:18:22 deriving 2012-04-18T06:18:29 (and it would have polluted namespace a bit) 2012-04-18T06:18:40 mleise: that is a new data type 2012-04-18T06:18:45 in brackets ... are functions used on that data type? 2012-04-18T06:18:52 and the stuff in the middle are fields? 2012-04-18T06:19:32 deriving (something) , means the compiler figures out automatically how to make that data type part of a typeclass 2012-04-18T06:19:43 Haskell is one of those things lots of people urge to have on your to-grok -list 2012-04-18T06:19:46 Oo it does _what_ ??? 2012-04-18T06:20:06 oh wait ... category theory? 2012-04-18T06:20:21 But not now :) 2012-04-18T06:20:32 mleise: Show means, you will be able to print the values of FColor, Eq means, you can compare two FColors, for equality, and Ord means, you can have ordering among them 2012-04-18T06:20:52 ok ok, so can I imagine that as a C++ interface? 2012-04-18T06:21:19 i think they sometimes compare typeclasses to java interfaces 2012-04-18T06:21:21 loosely 2012-04-18T06:21:53 keep in mind, that i could have given descriptive names to all those Double's 2012-04-18T06:21:55 like: 2012-04-18T06:22:14 data Ship = Ship {x :: Double, y :: Double,...} 2012-04-18T06:22:17 *** Garf has joined #aichallenge 2012-04-18T06:22:19 *** coeus has joined #aichallenge 2012-04-18T06:22:42 so later, i could have written: x aship 2012-04-18T06:22:44 ah, I'd really appreciate that in other people's code to understand the data types 2012-04-18T06:22:46 or: 2012-04-18T06:22:53 aship{y=34} 2012-04-18T06:24:17 mleise: data SO = SO FColor SceneObject deriving Show 2012-04-18T06:24:27 this is more interesting, when you can combine types 2012-04-18T06:25:04 btw, | means choices in type declarations 2012-04-18T06:25:15 ... 2012-04-18T06:25:25 data Int = 0 | 1 | 2 | ... 2012-04-18T06:25:34 you could define integers liek that 2012-04-18T06:25:49 0,1,2 is the data constructor here 2012-04-18T06:25:51 but there is only one type Int in a module, right? 2012-04-18T06:25:55 just like Ship, Rock 2012-04-18T06:26:03 yes 2012-04-18T06:26:16 you cant have name conflicts 2012-04-18T06:26:34 so SO = SO ... means ... class inheritance like in C++ ? 2012-04-18T06:27:07 hmm that makes no sense to me now that i think about it 2012-04-18T06:27:22 how does haskell know what SO is, when it wasn't previously defined 2012-04-18T06:27:33 no, it is a new type SO, which has a value constructor SO, and take 2 parameters, of type FCOlor and SceneObject 2012-04-18T06:27:59 SO is introduced here: data SO 2012-04-18T06:28:13 ah, so you repeat the type name there as the return value of the constructr 2012-04-18T06:28:31 btw, you dont have to structure your code, so that every definition of everything you use, must come before the code 2012-04-18T06:28:52 SO is the new type's name, and the contructor's too 2012-04-18T06:28:56 class SO { SO construct(FColor, SceneObject); } 2012-04-18T06:29:23 but it is not a class 2012-04-18T06:29:48 not even roughly ;) 2012-04-18T06:29:50 ? 2012-04-18T06:29:51 nothing good comes from trying to understand them from that perspective 2012-04-18T06:30:04 something will trip you up later, as they say 2012-04-18T06:30:10 oh come on, you have to fetch me from where I stand 2012-04-18T06:30:23 I only know classes and structs and all that stuff 2012-04-18T06:31:35 if you are interested read this http://www.haskell.org/haskellwiki/OOP_vs_type_classes 2012-04-18T06:31:58 i dont think my ideas about the language are really firm 2012-04-18T06:32:03 so i dont want to mislead you 2012-04-18T06:32:04 hmm, I'll give it a look 2012-04-18T06:32:15 though, i think im correct most of the time 2012-04-18T06:32:45 *** sigh has joined #aichallenge 2012-04-18T06:32:51 mleise: but you really need to distinguis between types and type classes 2012-04-18T06:33:18 sure I do. D has "ranges" for example 2012-04-18T06:33:25 whats that? 2012-04-18T06:33:33 they are no manifest type, but a set of operations available on any type 2012-04-18T06:33:59 so every type specializes a 'range' ? 2012-04-18T06:34:19 Actually they are just "if"s that check if a type has certain methods 2012-04-18T06:34:32 :) 2012-04-18T06:34:55 so an array offers the property .length, opIndex etc. 2012-04-18T06:35:21 and then it qualifies as a random access range by definition 2012-04-18T06:35:33 is that similar to type classes? 2012-04-18T06:37:34 well, not really 2012-04-18T06:38:00 type classes try to provide some operation on structures 2012-04-18T06:38:27 so when you implement the required operation for a new type,it will be part of that type class 2012-04-18T06:39:04 but it isnt about those kinds of abstractions, like providing random access 2012-04-18T06:40:24 "some operations" sounds to me just like that 2012-04-18T06:41:06 the functionality "Show" in your example allows the object to be printed by whatever mechanism there is in haskell 2012-04-18T06:41:31 and the functionality "length" and others allow a type to be used in interations and indexing in D 2012-04-18T06:42:30 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2012-04-18T06:42:34 So whatever the type is, it can be classified as "showable" or "indexable" 2012-04-18T06:42:58 yes, the Show typeclass requires a 'show' function to work on your data type, that must take your data, and make a string out of it 2012-04-18T06:43:04 but 2012-04-18T06:43:16 there is now typeclass that provides iteration for example 2012-04-18T06:43:38 usually, type classes used differently in haskell, than in other languages 2012-04-18T06:43:43 i mean in oop languages 2012-04-18T06:44:34 what all these typeclasses try to really do, is implement compositional properties for your types 2012-04-18T06:44:37 *** antimatroid has quit IRC (Ping timeout: 252 seconds) 2012-04-18T06:44:58 that's fine. i understand that approach 2012-04-18T06:45:09 *** delt0r_ has quit IRC (Ping timeout: 244 seconds) 2012-04-18T06:45:32 now if I give the array a method "toString()" it is a random access range and printable 2012-04-18T06:45:58 I can add more 'concepts' to the data type if I need them 2012-04-18T06:46:17 where the concepts would be either provided by the standard library or my own code 2012-04-18T06:48:02 I didn't like the way it was done in D much. I preferred strict interfaces. That makes me think: The "Show" in your code - is it a function name or some aggregate 'type' that can include a whole bunch of functions 2012-04-18T06:50:15 Show is the name of a typeclass and you need to provide the 'show' function for that 2012-04-18T06:50:38 basically, Show is parameterized over types 2012-04-18T06:50:39 ok, so it really is something I have to think of as a Java interface 2012-04-18T06:51:31 e.g. It adds an explicit requirement in terms of offered methods to a type. 2012-04-18T06:52:15 In D it is all implicit. You can have a typo in a method name and things may behave strange. 2012-04-18T06:52:33 *** alehorst1 has quit IRC (Quit: Leaving.) 2012-04-18T06:53:10 E.g. A random access range becomes a forward range and cannot be indexed any more, resulting in more or less comprehensable error messages when you use algorithms that need indexing. 2012-04-18T06:54:43 in haskell you must provide a minimal generator set of functions for your typeclass 2012-04-18T06:54:49 with the proper signature 2012-04-18T06:56:12 minimal generator set means what ? comparison and arithmetics? 2012-04-18T06:56:24 no 2012-04-18T06:56:29 type construction? 2012-04-18T06:57:00 means a typeclass, can require different functions for it to work, and it can define those in terms of others 2012-04-18T06:57:19 and you only need to provide so much for it to work, not all of them 2012-04-18T06:57:25 so many* 2012-04-18T06:57:28 *** delt0r_ has joined #aichallenge 2012-04-18T06:57:55 an example is == and /= 2012-04-18T06:58:06 *** antimatroid has joined #aichallenge 2012-04-18T06:58:28 x==y = not x/=y 2012-04-18T06:58:38 x/=y = not x==y 2012-04-18T06:58:53 these are the two operations the Eq typeclass needs 2012-04-18T06:58:59 Ok, that's compiler magic that works in many languages 2012-04-18T06:59:04 and you only need to provide == or /= 2012-04-18T06:59:09 no 2012-04-18T06:59:19 what compiler magic? 2012-04-18T06:59:34 The compiler knows internally that equals is the inverse of not equals 2012-04-18T06:59:53 or does the class Eq define these operations 2012-04-18T06:59:55 i just stated it, the compilers no jackshit about it 2012-04-18T06:59:59 knows* 2012-04-18T07:00:27 == and /= are no dfferent from someFunction 2012-04-18T07:00:53 or .\=* which you call an operator, but is like a function ike any other 2012-04-18T07:01:15 in haskell there is only a tiny distinction between operators and functions 2012-04-18T07:01:40 and the typeclass Eq provides the information that you only need to define either == or /= for it to work? 2012-04-18T07:01:59 symbols made from non alphanumeric characters are treated as infix, and others are prefix 2012-04-18T07:02:16 class Eq a where 2012-04-18T07:02:31 (==) :: a -> a -> Bool 2012-04-18T07:02:33 (/=) :: a -> a -> Bool 2012-04-18T07:02:39 x==y = not x/=y 2012-04-18T07:02:43 x/=y = not x==y 2012-04-18T07:02:52 this is the definition of the typeclass Eq 2012-04-18T07:03:17 that's neat. now I understand in part why this system is good 2012-04-18T07:03:56 a here, is a type variable 2012-04-18T07:04:02 like a template parameter 2012-04-18T07:04:48 "on any a offer a method that takes an a and returns a bool" 2012-04-18T07:05:29 again, 'a' does not offer a method 2012-04-18T07:05:59 :( 2012-04-18T07:06:31 maybe I'll try to convert your pasted code snippet to D to understand what it really does 2012-04-18T07:06:49 ^^ this? 2012-04-18T07:07:07 the one in that haskell paste site of course ;) 2012-04-18T07:07:12 ah 2012-04-18T07:07:19 is it a snippet? 2012-04-18T07:07:25 it is a complete working program 2012-04-18T07:07:45 do you really want to translate it? 2012-04-18T07:07:53 it would be fun 2012-04-18T07:08:04 does D have gtk bindings? 2012-04-18T07:08:08 XD 2012-04-18T07:08:38 yes it does, good ones even! 2012-04-18T07:08:42 im kidding 2012-04-18T07:08:49 everything has gtk bindings 2012-04-18T07:09:03 i dont know how good haskell's is 2012-04-18T07:09:07 well, it needs someone to write them 2012-04-18T07:09:12 it was easy, but i really cant compare 2012-04-18T07:09:17 (im a qt guy) 2012-04-18T07:09:19 there are mediocre qt4 bindings for example 2012-04-18T07:09:29 and not so good wxWidgets bindings 2012-04-18T07:09:46 mleise: but do you really want to translate that code? 2012-04-18T07:09:51 or you were just kidding 2012-04-18T07:09:52 Gtk+ is the only GUI toolkit that currently just works 2012-04-18T07:10:21 no I really want to translate that code to see what the constructs in Haskell do and what I miss or don't miss out on 2012-04-18T07:10:35 but that is just a beginner's code 2012-04-18T07:10:39 and it is a first pass 2012-04-18T07:10:54 i mean i didnt rewrite anything 2012-04-18T07:10:55 I'm not an expert in D either 2012-04-18T07:11:06 come on, you must be a much better programmer than me 2012-04-18T07:11:16 i know you are 2012-04-18T07:11:37 lol, how do you think that? I don't even know partial function application 2012-04-18T07:11:48 but I sure know some ASM 2012-04-18T07:11:49 but you write code for aliving 2012-04-18T07:12:03 Yes, and you will soon enough I think 2012-04-18T07:12:12 well, i do per say 2012-04-18T07:12:19 but this is a tiny project 2012-04-18T07:12:24 (i mean the Qt one) 2012-04-18T07:13:10 Actually it only ever used Java and Delphi in the business world and a bit of ASP.net 2012-04-18T07:13:25 s/it/I/ 2012-04-18T07:13:58 you know the time when you talk about yourself in the third person ... 2012-04-18T07:14:14 that code is under 200 lines, and i wrote it in 8 hours, i think im a lousy programmer 2012-04-18T07:14:32 though, i had to dig in the documentations 2012-04-18T07:14:49 mleise: yes, sign of insanity 2012-04-18T07:14:55 * mcstar likes that 2012-04-18T07:15:15 sure, I'm not particularly fast either. I cannot type down a program without stopping 2012-04-18T07:15:31 mleise: i think you can do it in 100 lines 2012-04-18T07:15:39 I'm always stopping to think if what I just wrote makes sense 2012-04-18T07:15:50 100 lines of haskell 2012-04-18T07:15:54 no, D 2012-04-18T07:15:57 Oo 2012-04-18T07:16:12 you know we use { and } and the likes ? 2012-04-18T07:16:13 look, my imports alone 12 lines 2012-04-18T07:16:32 mleise: only counting meaningful ines 2012-04-18T07:16:35 lines* 2012-04-18T07:16:55 a struct definition has every field on a separate line as well 2012-04-18T07:17:08 i think it will be much longer 2012-04-18T07:17:13 *** antimatroid has quit IRC (Ping timeout: 264 seconds) 2012-04-18T07:17:19 anyway, lets see 2012-04-18T07:17:24 than we cmpare speeds 2012-04-18T07:17:34 oh ... you don't really want that 2012-04-18T07:17:51 why? are you too good in optimizations? 2012-04-18T07:18:15 I claim to be good at it, but the GCC is even better at it ;) 2012-04-18T07:18:37 it will be 10 times faster than the haskell code 2012-04-18T07:18:39 ghc is good at it too 2012-04-18T07:18:42 oh 2012-04-18T07:18:50 well, maybe ... 2 times 2012-04-18T07:18:59 yeah, thats more reasonable 2012-04-18T07:19:12 anyway time for lunch 2012-04-18T07:19:30 mleise: when will you do it? 2012-04-18T07:19:47 this evening 2012-04-18T07:20:05 ill be around, so you can aske me, if you need to 2012-04-18T07:20:06 hey, you could have commented your code 2012-04-18T07:20:13 i dont comment, ever 2012-04-18T07:20:52 you should have seen my Ants entry... 2012-04-18T07:24:25 0 comments 2012-04-18T07:25:20 you too? 2012-04-18T07:25:26 Pretty much, yea 2012-04-18T07:25:30 hehe, well. when you work for money you should leave comments 2012-04-18T07:25:45 it helps others in the team and thos who come after you 2012-04-18T07:26:08 the first thing you will notice is the absence of good comments when you start a at a new job 2012-04-18T07:26:27 *noone* writes them, but *everyone* misses them 2012-04-18T07:27:07 for ants it may be ok, because you neither reuse or share the code 2012-04-18T07:27:22 well ... actually you did share it, but you get the point 2012-04-18T07:27:46 i think comments would just get in the way, and destroy readability 2012-04-18T07:27:57 the problem was, that my code was lame/unstructured 2012-04-18T07:28:33 My code was uhh, let's not talk about my code 2012-04-18T07:28:43 yeah, i second that 2012-04-18T07:29:54 I only worked on it for the first ~week of the competition 2012-04-18T07:34:15 *** antimatroid has joined #aichallenge 2012-04-18T07:34:28 *** GarfTop has joined #aichallenge 2012-04-18T09:02:53 *** amstan has joined #aichallenge 2012-04-18T09:02:53 *** ChanServ sets mode: +o amstan 2012-04-18T10:06:33 *** coeus has quit IRC (Quit: Verlassend) 2012-04-18T10:32:40 *** NoxiaZ^ has joined #aichallenge 2012-04-18T10:32:53 *** kurnevsky has joined #aichallenge 2012-04-18T10:37:18 *** Areks has quit IRC (Ping timeout: 272 seconds) 2012-04-18T10:38:27 *** thestinger has joined #aichallenge 2012-04-18T10:42:50 mcstar: I tried your collision thing btw 2012-04-18T10:42:57 I only used cabal for mersenne-random though 2012-04-18T10:43:14 you had gtk already> 2012-04-18T10:43:17 ? 2012-04-18T10:43:45 no, used the arch linux package :P 2012-04-18T10:44:02 i decided not to install things that way 2012-04-18T10:44:43 thestinger: mleise decided to rewrite it in D 2012-04-18T10:44:58 he says he hates haskell, and really cant stand the code 2012-04-18T10:46:15 doing cs212 (taught by peter norvig) and udacity's crypto course 2012-04-18T10:46:42 yeah, for example, what does "where a b = someFunc ..." mean ? 2012-04-18T10:47:04 how can anyone understand this pile of crap? 2012-04-18T10:47:12 it has no clear structure or anything 2012-04-18T10:47:19 yes it does 2012-04-18T10:47:33 indentation is significant, like python 2012-04-18T10:47:42 you can choose to use braces but no one does 2012-04-18T10:47:53 but this line in mcstar's code is not indented 2012-04-18T10:48:36 which? 2012-04-18T10:48:38 it's basically just a local function definition (for that block) 2012-04-18T10:48:39 it is the last in a sequence of "do" operations 2012-04-18T10:48:49 mleise: line number? 2012-04-18T10:48:55 180 2012-04-18T10:49:16 I understand the glib part, but not the haskell part 2012-04-18T10:49:42 mleise: it's a local function definition (for the block above) 2012-04-18T10:49:43 cs 253 web app engineering's only homework was to set-up google app engine 2012-04-18T10:49:44 form definitions follow after where 2012-04-18T10:49:52 it is quite self explanatory 2012-04-18T10:50:05 form definitions? 2012-04-18T10:50:08 yes 2012-04-18T10:50:15 something = 2012-04-18T10:50:19 let's start with the block above 2012-04-18T10:50:22 something some other = 2012-04-18T10:50:42 doPeriodic $ do (w, h) <- ... 2012-04-18T10:50:46 something can be a substitution, or it can have parameters 2012-04-18T10:50:51 (you call it function) 2012-04-18T10:50:53 is that a proper inner function 2012-04-18T10:51:03 it is an action 2012-04-18T10:51:15 it is not a lambda or function 2012-04-18T10:51:34 but that isnt like 'where' 2012-04-18T10:51:36 so ... just a block of code? 2012-04-18T10:52:04 it must have a proper type 2012-04-18T10:52:17 what is the type of doPeriodic? 2012-04-18T10:52:21 it has IO (Bool) 2012-04-18T10:52:47 you can see, that it passes that action to timeoutAdd 2012-04-18T10:52:54 I can't make out where the "$" ends, i.e. where the ) bracket would be if you wrote ( ...) instead of $ 2012-04-18T10:53:11 just follow indentation 2012-04-18T10:53:41 $ is just there, so that i dont have to put in () the whole action 2012-04-18T10:54:02 the *whole* action, so that means down to line 77 2012-04-18T10:54:04 177 2012-04-18T10:54:09 yes 2012-04-18T10:54:57 ok, understood. "doPeriodic ( ... ) " defines a new action. let us call it a closure with no arguments 2012-04-18T10:54:58 mcstar: seems like I have an older version of your code 2012-04-18T10:55:29 thestinger: technically, the difference is the window size and refreash rate 2012-04-18T10:55:44 mcstar: you may want to use Gtk's MainWindow. The application automatically terminates when the last MainWindow is closed 2012-04-18T10:55:46 mleise: no 2012-04-18T10:55:51 yes 2012-04-18T10:55:58 doPeriodic does not define a new action 2012-04-18T10:56:24 but ... it is a block of code with a name, right? 2012-04-18T10:56:28 mleise: do mcstar is teaching you haskell? :P 2012-04-18T10:56:28 no 2012-04-18T10:56:38 is .. teching 2012-04-18T10:56:55 weird, the algo-class stopped posting new content 2012-04-18T10:56:58 :\ 2012-04-18T10:57:04 mleise: 'do something' introduces an action 2012-04-18T10:57:13 doPeriodic takes an action as parameter 2012-04-18T10:57:21 thestinger: yes, he doesn't really want to, but with his own code as the basis it works :p 2012-04-18T10:57:35 wait, you lost me there 2012-04-18T10:57:55 where is that action parameter visible? 2012-04-18T10:58:04 i only see "doPeriodic ( ... )" 2012-04-18T10:58:11 doPeriodic action = timeoutAdd action 10 2012-04-18T10:58:32 but ... above you defined it as just a squence of commands 2012-04-18T10:58:45 how can it suddenly take a parameter 2012-04-18T10:58:52 is it a template? 2012-04-18T10:59:10 the action, the argument to doPeriodic is a sequence of other actions 2012-04-18T10:59:29 mcstar: http://ompldr.org/vZGY2aw basically the same stuff we did in week1 of the coursera crypto class :P 2012-04-18T10:59:36 nice way to review it 2012-04-18T11:00:01 2-3 minute videos with a quiz between each is easier to digest than big 30 min lectures with 1 quiz 2012-04-18T11:00:54 mleise: doPeriodic is just like any other function, here it is just a shorthand for timeoutAdd 2012-04-18T11:00:55 mcstar: you said doPeriodic is an action, now it takes an action as a parameter 2012-04-18T11:00:56 I expect the homework won't be nearly as challenging as the coursera stuff though 2012-04-18T11:01:19 *** kilae has quit IRC (Quit: ChatZilla 0.9.88.2 [Firefox 11.0/20120312181643]) 2012-04-18T11:01:33 mleise: whats wrong with that? 2012-04-18T11:01:46 functions take functions as parameters too 2012-04-18T11:01:56 *is* it an action or does it *take* an action 2012-04-18T11:02:09 you call something an action, cause it has a bit different type signature, it has -> IO (a) 2012-04-18T11:02:22 since it takes an action, it will become one too 2012-04-18T11:02:32 ok, so this is more of an internal distinction 2012-04-18T11:02:42 to allow for non-pure code 2012-04-18T11:02:49 that works in sequence 2012-04-18T11:02:57 etc. 2012-04-18T11:03:00 here the order matters 2012-04-18T11:03:03 not pureness 2012-04-18T11:03:44 the 'return type' of doPeriodic is determined by the last action, i.e. return True 2012-04-18T11:04:53 Finished with week1 of the 4 udacity courses in a bit 2012-04-18T11:05:13 even in these cases, you can just use substitution as the programming model, to understand whats going on 2012-04-18T11:05:19 the problem is with IORef 2012-04-18T11:05:28 I'm still not quite sure what doPeriodic is now. It doesn't seem to take any parameters (line 157), but you say it takes an action as the parameter 2012-04-18T11:05:29 since that really introduces mutable state 2012-04-18T11:05:29 pairofdice: how are the two I haven't signed up for? :P (I'm taking the cs212 and crypto atm) 2012-04-18T11:05:44 crypto will be review now... and I don't think cs212 will be hard 2012-04-18T11:05:54 the homework was one liners :P 2012-04-18T11:06:06 mleise: wha do you mean? everything that comes after $ is the parameter of doPeriodic 2012-04-18T11:06:07 programming languages had reg exp, cs253 basics of html 2012-04-18T11:06:11 and http 2012-04-18T11:06:20 might as well take them 2012-04-18T11:06:24 ok big misunderstanding then :p 2012-04-18T11:06:27 I hope they end up doing harder courses 2012-04-18T11:07:01 mleise: doPeriodic $ do this; that; return some == doPeriodic (do this; that; return some) 2012-04-18T11:07:16 pairofdice: the format of udacity's stuff is _way_ better than the coursera ones 2012-04-18T11:07:18 but the depth isn't there 2012-04-18T11:07:37 Yeah, I guess they're building up to more challenging stuff 2012-04-18T11:07:51 what did I actually learn in cs373? not very much 2012-04-18T11:08:11 I learned a lot more about pathfinding in graphs in the 1 week that the algo-class covered it 2012-04-18T11:08:17